THE MELILITE-BEARING HIGH-TEMPERATURE SKARNS …rruff.info/doclib/cm/vol39/CM39_1405.pdf · MELILITE-BEARING HIGH-TEMPERATURE SKARNS, ROMANIA 1407 (mainly granite porphyry and granophyre),

1405

§ E-mail address: [email protected]

The Canadian MineralogistVol. 39, pp. 1405-1434 (2001)

THE MELILITE-BEARING HIGH-TEMPERATURE SKARNSOF THE APUSENI MOUNTAINS, CARPATHIANS, ROMANIA

MARIE-LOLA PASCAL§

CNRS–ISTO (Institut des Sciences de la Terre d’Orléans), 1A, rue de la Férollerie, F-45071 Orléans Cedex 02, France

MICHEL FONTEILLES

Laboratoire de Pétrologie, Université Pierre-et-Marie-Curie, 4, place Jussieu, F-75252 Paris Cedex 05, France

JEAN VERKAEREN AND RÉGIS PIRET

Unité de Géologie, Université Catholique de Louvain-la-Neuve, Batiment Mercator, 3, place Louis Pasteur,B-1348 Louvain-la-Neuve, Belgium

STEFAN MARINCEA

Geological Institute of Romania, 1, Caransebes Street, RO-78344 Bucharest, Romania

ABSTRACT

The melilite-bearing skarns of Cornet Hill (CH) and Upper Cerboaia Valley (UCV), in the Apuseni Mountains of Romania,occur at the contact between monzodiorite bodies of Ypresian age (Paleocene) and Neojurassic calcitic marbles. Typicalwollastonite – grossular – diopside endoskarns are separated from exoskarns (tilleyite and spurrite or wollastonite at CH,wollastonite only at UCV), at most places, by a melilite-rich rock, in which veins and vein-like zones of recrystallization arecomposed only of idiomorphic melilite crystals reaching 15 cm across. Titanian garnet and wollastonite are the principal mineralsassociated with melilite (and also monticellite, perovskite, vesuvianite, cuspidine, spurrite, tilleyite, calcite, hydroxylellestadite,hydrogrossular and other minor alteration-induced minerals). A different association that includes aluminian diopside andgrossular occurs (1) as veinlets in the marble close to the skarns and (2) as relict inclusions in endoskarns. From the geometricalrelationships of the zone sequences and the veins, the textural features of the mineral associations and the inferred conditions offluid–mineral equilibrium, these mineralogical peculiarities are interpreted as resulting from the superposition of two main stages.Firstly, there was circulation of a comparatively CO2-rich fluid formed the early aluminian diopside – grossular endoskarns, withdepletion in Si (and Fe, Na, K) and inert behavior of Mg, Al, Ti. Then, a high-temperature (750°C) fluid circulated on both sidesof the contact between marble and endoskarns, and developed the melilite-rich, titanian-garnet-bearing rocks partly at the expenseof previously formed endoskarns, and spurrite or wollastonite (CH) or wollastonite (UCV) exoskarns at the expense of marble.The pressure of CO2 was very low, less than 26 bars at UCV and 16 bars at CH, with a H2O pressure less than 750 bars. Not onlySi and Ca were mobilized, but also Mg, Al and Ti, leached from the endoskarns and deposited in the veins and the nearby part ofexoskarns. This stage, which occurred in the temperature range corresponding to the end of the crystallization of plagioclase inthe monzodiorite, has pegmatitic chemical and textural features. The main flow of fluid ended with the development of tilleyitepartly at the expense of spurrite and wollastonite at CH, and local high-temperature (about 710°C) recrystallization of the zonation,mostly in veins, especially in the endoskarn–exoskarn boundary, but also within the endoskarns. A monticellite – gehleniteassociation appeared in the melilite-rich rocks, later followed by vesuvianite, whereas in the endoskarn, vesuvianite developedtogether with coarse-grained wollastonite and grossular.

Keywords: skarns, melilite, spurrite, tilleyite, titanian garnet, metasomatism, Apuseni Mountains, Romania.

SOMMAIRE

Les skarns à mélilite de Cornet Hill (CH) et de la haute vallée de la Cerboaia (UCV), dans les monts Apuseni, en Roumanie,sont développés au contact de corps intrusifs monzodioritiques, d’âge Ypresien (Paléocène), aux dépens de ceux-ci et des marbresencaissants, purement calcitiques, d’âge néojurassique. Les endoskarns typiques, caractérisés par l’association wollastonite –grossulaire – diopside, sont séparés des exoskarns (wollastonite à UCV, tilleyite et spurrite ou wollastonite à CH) par une roche

1405 39#5-oct-01-2277-15 26/10/01, 13:011405

1406 THE CANADIAN MINERALOGIST

INTRODUCTION

The skarn occurrences to be described here are lo-cated in the Apuseni Mountains of Romania, about 20km west of the town of Brad. Istrate et al. (1978), Stefanet al. (1978) and Piret (1997) have reported the pres-ence of high-temperature minerals such as melilite,spurrite and tilleyite in these skarns. Few occurrencesof such minerals have been described in literature. Someare products of high-grade metamorphism of siliceouslimestone [e.g., Carlingford, Ireland: Nockolds (1947),Camas Mor, Scotland: Tilley (1948, 1951), TunguskaRiver basin, Siberia: Reverdatto (1970), Reverdatto etal. (1979), and Hatrurim, Israel: Gross (1977)], or ofcalcic enclaves in melts [Oslo Rift: Jamtveit et al. (1992,1997), Willemse & Bensch (1964)]. Others are skarnassociations [e.g., Crestmore, California: Burnham(1959), Wiechmann (1995), Scawt Hill, Ireland: Tilley(1929), Carneal, Ireland: Sabine (1975), Isle of Rhum,Scotland: Hughes (1960), Christmas Mountains, Texas:Joesten (1974), Joesten & Fisher (1988), Kushiro, Ja-pan (Henmi et al. 1971), and Fuka, Japan (Henmi et al.1973)]. Of these, the Crestmore occurrence has been thefocus of particular attention by mineralogists and pe-trologists and constitutes a reference for this peculiartype of skarn.

The classical interpretation of skarn formation in-volves (i) the infiltration of a magmatic fluid into thesurrounding rocks and (ii) the development of a se-quence of continuous zones tending to be monomi-neralic, according to the chromatographic model(Korzhinskii 1970). Where skarn veins are observed,they should represent the most thoroughly fluid-con-trolled part of the system, and the zone sequences areexpected to be organized around them. However, in theskarns at Crestmore, the associations described byBurnham (1959) in the monticellite zone show neither a

continuous development nor a tendency to a monomi-neralic character. These skarns thus cannot be com-pletely described as a chromatographic sequence. Theexistence in skarns of hydrothermal processes other thanthe outward infiltration of a magmatic fluid is also indi-cated by available isotopic signatures (O, H) of skarnminerals, which commonly indicate a dominantly meta-morphic origin for the fluid (Guy et al. 1988). In addi-tion, a large-scale development of typical endoskarns iseasier to account for in the case of a (locally) externalorigin for part of the fluids.

The high temperature that characterizes the mineralassociations of the skarns of the Apuseni Mountainsindicates that products of the early metasomatic pro-cesses have not been extensively reworked by late hy-drothermal circulation; therefore, these skarns areappropriate to illustrate those features than can be at-tributed to the chromatographic model from those thatcannot. As the veins and their geometrical and chrono-logical relations to the zonal pattern offer clues relevantto this issue, they have been described with the samedetail as the zonal sequences, in order to decide whichminerals, with which compositions, were associated atequilibrium.

GEOLOGICAL SETTING AND FIELD RELATIONSHIPS

In the region of Magureaua Vatei (Fig. 1), the geo-logical formations include a thick (>3000 m) sequenceof calc-alkaline mafic volcanic rocks and gabbros, the“ophiolites” of Drocea, of Upper Triassic to Jurassicage. Neojurassic limestones follow, overlain by Creta-ceous clastic sediments and Upper Mesozoic basalts andandesites. This sequence is cut by felsic hypabyssalbodies of Ypresian age (Paleocene), the so-calledbanatites, ranging in composition from dominantmonzodiorite to minor monzonite, syenite and granite

à mélilite dominante, d’épaisseur variable, localement absente, comportant des veines et des recristallisations uniquement forméesde mélilite en cristaux idiomorphes atteignant 15 cm. Les principaux minéraux associés à la mélilite sont le grenat titanifère et lawollastonite, auxquels s’ajoutent monticellite, pérovskite, vésuvianite, cuspidine, spurrite, tilleyite, calcite, hydroxylellestadite,hydrogrossulaire et divers minéraux d’altération. Une paragénèse différente, comportant diopside alumineux et grossulaire,apparait soit en veinules dans le marbre à proximité immédiate des skarns, soit en reliques dans l’endoskarn. A partir des relationsgéométriques entre séquences de zones et veines, des compositions des minéraux et de leurs relations texturales, et de lamodélisation thermochimique de leurs conditions de stabilité, ces particularités sont interprétées par la superposition de deuxprocessus hydrothermaux de haute température et basse pression [750°C, P(H2O) < 750 bars], chimiquement contrastés: 1)formation d’endoskarns précoces à diopside alumineux – grossulaire, par un fluide relativement riche en CO2 [P(CO2) > 100 bars]qui lessive Si, Fe, Na, K sans mobiliser Mg, Al et Ti. 2) Circulation, le long du contact entre marbres et endoskarns, d’un fluidecaractérisé par une très faible pression de CO2 (moins de 26 bars à UCV, moins de 16 bars à CH), qui développe la mélilite et legrenat titanifère en veines et aux dépens des endoskarns, et transforme le marbre en spurrite ou wollastonite à CH, en wollastoniteà UCV. Les transformations observées impliquent un apport de Mg, Al, Si et Ti dans les exoskarns proches du contact, et unlessivage de Mg et Ti dans les endoskarns. Par ses particularités chimiques et texturales et en accord avec sa température, quicorrespond à la fin de cristallisation de la monzodiorite, ce processus présente un caractère pegmatitique. Ce stade évolue avec ledéveloppement de la tilleyite à CH, en partie aux dépens de la spurrite et de la wollastonite, et prend fin vers 710°C avec desmodifications secondaires locales, principalement en veines: dans les roches à mélilite apparaissent une association à monticellite– gehlenite, puis la vésuvianite, et dans l’endoskarn, la vésuvianite et des recristallisations de la wollastonite et du grenat.

Mots-clés: skarns, mélilite, spurrite, tilleyite, grenat titanifère, métasomatose, monts Apuseni, Roumanie.

1405 39#5-oct-01-2277-15 26/10/01, 13:011406

MELILITE-BEARING HIGH-TEMPERATURE SKARNS, ROMANIA 1407

(mainly granite porphyry and granophyre), at the con-tact of which the high-temperature skarns have beenobserved. Two occurrences are described here: (1) avery small one on top of Cornet Hill (CH), where thecontact outcrops over twenty meters, and (2) an evensmaller outcrop of an intrusive contact in the UpperCerboaia Valley (UCV).

In both occurrences, the host marble is exclusivelycomposed of pure calcite. At any one place, the originalcontact is masked by several sequences of mineral as-sociations, of which 52 samples were collected. Thesequences mainly differ by the mineralogical nature ofthe exoskarns, spurrite – tilleyite ± wollastonite at CH,wollastonite ± calcite at UCV. Two sequences are de-scribed below; sequence S is from Cornet Hill, and se-quence W is from the Upper Cerboaia Valley.

Sequence S

Over a few meters at the contact with the quartzmonzodiorite, Istrate et al. (1978) reported the follow-ing zonation: quartz monzodiorite / wollastonite +melilite + vesuvianite / spurrite / tilleyite / calcite. Ac-cording to our observations, the wollastonite – melilite– vesuvianite zone of Istrate et al. (1978) includes twodistinct parts. One is composed of wollastonite – gros-sular ± pyroxene (zone A, Table 1), and the other, ofmelilite as the dominant mineral (zone B). The spurriteand tilleyite zones of Istrate et al. (1978) (C and D),mainly composed of very coarse-grained carbonate-bearing minerals (with some wollastonite, melilite,titanian garnet, and vesuvianite), are interpreted asexoskarns. According to the field observations of Istrateet al. (1978), the zones are reported to be several meterswide, but as they occur as a back-slope, with rather poorconditions of outcrop, the thicknesses may have beenoverestimated; no transverse section exists through thissystem of zones.

The succession of mineral associations shown inTable 1 is that obtained after eliminating secondarymodifications and alterations, which are examined indetail in a companion paper by Marincea et al. (2001).The specific minerals are listed with their structural for-mulae in Table 2. As spurrite, of major petrological sig-nificance, has been observed only in a few samples,Table 1 shows the spurrite-free sequence as the mainone, and the location of spurrite-bearing zones as vari-ants of the sequence. Most skarn samples can be locatedin this sequence of zones. One sample (CH22) shows awollastonite exoskarn zone without melilite, theexoskarn being in direct contact with typical grossular– wollastonite endoskarns (zone A).

Sequence W

The only difference between sequence W and se-quence S is the absence of spurrite and tilleyite. Theexoskarn (zone C), composed of calcite – wollastonite

(with some melilite and titanian garnet), is in contactwith zone B, containing the assemblage melilite – wol-lastonite – titanian garnet.

Melilite veins

The boundary between zones B and C is occupied bya vein (“BV”) a few mm thick at CH and up to 1 m atUCV, composed of coarse melilite (up to 0.5 cm at CH,more than 15 cm at UCV) (Plates 1B and 2 in Stefan etal. 1978). Many similar veinlets of melilite (in most casesassociated with titanian garnet), millimeters in thicknessat CH and centimeters at UCV, have been observed inexoskarns, branching off the main vein (Fig. 2).

Pyroxene veins

In the marble at UCV, very close to the melilite –wollastonite rock and one meter away from the quartzmonzonite, a complicated and dense network of vein-lets, at most a few mm thick, is mainly composed ofpyroxene and garnet (sample UCV4–1).

PETROGRAPHY OF THE INTRUSIVE ROCKS

Monzodiorites

The primary minerals present in the monzodioriteincludes abundant plagioclase (An40–60, An25 at the rim),some orthopyroxene or amphibole (or both), muchclinopyroxene and biotite, magnetite and ilmenite, K-feldspar and quartz. Accessory minerals are apatite andscarce zircon. The most important difference betweenCH and UCV localities is the occurrence at CH of smallearly phenocrysts of greyish Ti-bearing edenite, whereasat UCV a brownish Ti-rich edenite is only observed asinclusions in the center of zoned grains of clino-pyroxene.

K-feldspar shows a tendency to concentrate inpatches (monzonitic texture). In many samples, at leastpart of the quartz appears to be late to develop, mainlyat the expense of K-feldspar. Interstitial granophyre isstill locally observed, and may constitute up to 80% ofthe rock, as in sample UCV3, in which the other con-stituents are the same as in the normal monzodiorite(e.g., same composition of orthopyroxene). Theamounts of K-feldspar and quartz are highly variable,even at the scale of the hand specimen (UCV63), sug-gesting an appreciable mobility of the pegmatite-form-ing melt with respect to the already crystallized dioriticframework of the rock. This melt, either granitic or sy-enitic, shows a tendency to concentrate in the marginsof the intrusive body.

Contaminated rocks and “syenite”

Close to the contact with the skarns, some changesin the associations of the intrusive rocks, especially in

1405 39#5-oct-01-2277-15 26/10/01, 13:011407


their pegmatitic parts, correspond to what is commonlyattributed to contamination. At the contact with theendoskarns, the rock is referred to as “syenite”. Thesechanges show a more or less clearly defined zonal pat-tern, with at first the disappearance of ilmenite and theappearance of titanite. The corrosion of plagioclase isaccompanied by the development of an association ofK-feldspar, quartz and greenish clinopyroxene. Fine-grained polycrystalline aggregates of clinopyroxene,biotite and magnetite are pseudomorphic after ortho-pyroxene and hornblende, which completely disappear.

The next step is the disappearance of primary biotite,accompanied by the further development of clino-pyroxene and, in syenitic varieties, by the formation ofnepheline, interpreted as due to this silica-consumingreaction. Sample CH2, composed of large crystals of

nonperthitic and fresh alkali feldspar (Or64–68) that in-clude a multitude of unoriented very small idiomorphiccrystals of oligoclase (An22–26) and diopside (Di80), isconsidered to represent the last stage of plagioclase sta-bility, before a transition to typical alkali feldsparpegmatites, which occur as large veins in this rock. Suchpegmatites include quartz-bearing and nepheline-bear-ing varieties. In some of them, the coarse-grained or-thoclase (12–20% Ab, 0.20–0.37% An and 0.09–0.12wt.% FeO) leaves large intercrystalline spaces filledwith hydrothermal material (wollastonite, calcite,prehnite, pectolite, analcime). The latest step of pegma-tite evolution corresponds to the development of brightgreen fringes, apophyses and cross-cutting veinlets, witha more-or-less aegirine-rich composition, on the pre-existing, commonly idiomorphic crystals of clino-pyroxene. At their outermost rim, a brown colorobserved at some places corresponds to Ti-bearingaegirine. The only prominent accessory mineral istitanite.

Within the border region, the last change observed,at a few centimeters of the typical endoskarn boundary,is the disappearance of magnetite along a sharp front.

Thermobarometry

Crystallization temperatures are provided by thecompositions of coexisting plagioclase and K-feldspar(CH2, UCV21). Assuming that the An content of thetwo feldspars has not re-equilibrated, temperatures werecalculated according to the model for ternary feldsparsof Green & Usdansky (1986) (Fig. 3). Good agreementis observed between measured and calculated composi-tions of coexisting phases, except for the Or content ofthe plagioclase, which is systematically lower than an-ticipated from calculation, most probably as a result oflater K–Na exchange with the surrounding K-feldspar(Table 3). Such later exchange is not surprising, owingto the very small size of the plagioclase inclusions im-bedded in K-feldspar in the “syenite” (sample CH2). Inthe “syenite”, the observed pairs of coexisting feldsparscrystallized from 820°C down to 760°C, and mostlyaround 790–800°C. In the case of the more internal part

1405 39#5-oct-01-2277-15 26/10/01, 13:011408


of the intrusion at UCV, crystallization temperatures inthe range 800–740°C have been obtained.

In this temperature range, hornblende is stable at anH2O pressure down to 0.5 kilobar (Burnham 1979).Whereas this mineral has been observed as early phe-nocrysts in the monzodiorite at CH, it only appears invery small amount at UCV, as probable remnants, al-most entirely transformed to clinopyroxene. Themonzodiorite plutons were thus emplaced in a shallowenvironment, with a H2O pressure slightly higher at CHthan at UCV, both being around 0.5 kilobar.

PARAGENETIC DESCRIPTION OF THE ZONED SKARNS

The changes in the intrusive rock are described fromthe endoskarn – “syenite” boundary outward, and willbe followed by a description of the melilite veins andexoskarns. Zone A is composed of several subzones thatmay or may not be present in a given sample.

Zone A1

The endoskarns are demarcated by a very clear con-tact-parallel plane, which constitutes the first skarn front

FIG. 1. Geological sketch of the Cerboaia Valley region (redrawn from Stefan et al. 1978).

1405 39#5-oct-01-2277-15 26/10/01, 13:011409


in contact with the “syenite” (Fig. 4A). In the intrusiverock along the contact, a calcic mineral with a zeolite-like organization and a composition Na0.10Ca2Al1.10Si3O9.70(H2O) was observed, with a thickness up to5 mm.

The pyroxene is idiomorphic and zoned, with thesame grain-size and overall shape as in the intrusiverock, but with a different aspect (CH78). The rim pro-gressively loses its shape and color, and becomes sieve-like. In other cases, a rim of fibrous wollastonite withfibers parallel to the pyroxene boundary develops at itsexpense. In some instances, it shows outer zones of in-tense green color, but interrupted and surrounded bywollastonite. Its composition (Table 4) shows changesin Fe and Na contents from the typical values in the“syenite” (1–10 wt.% FeO, 0.1–1.6% Na2O, 4.5–0.4%Al2O3), to values that compare with those in zone A2(less than 2% FeO, less than 0.1% Na2O, 1.5–6%Al2O3).

Wollastonite is abundant and mainly occurs assubparallel fibers perpendicular to the contact, orga-

nized in millimeter-thick layers separated by surfacesparallel to the contact with the intrusive rock (Fig. 4A).It is associated with very fine-grained grossular, andboth conspicuously developed at the expense of K-feld-spar. The grain size increases very progressively acrossthese surfaces from the contact with the intrusive rock(less than 1 �m) outward.

Perovskite may occur, on the side of the intrusiverock, as very fine-grained, opaque or nearly opaqueaggregates of octahedra, replacing the large crystals oftitanite observed in the “syenite” (Fig. 4A).

A mineral association intermediate between that justdescribed and the “syenite”, with K-feldspar and titaniteinstead of grossular and perovskite, has been observedas a zone 1 to 2 mm thick in two samples.

Zone A2

Zone A2 is characterized by changes in aspect andcomposition of pyroxene crystals, which progressivelylose their hypidiomorphic outline, leading to irregularlyshaped aggregates of interlocking crystals of diopside(XMg in the range 0.90–0.98). Exceptionally, the largergrains show a core with a composition typical of the“syenite”. Corroded grains or aggregates with a distri-

FIG. 3. An contents of coexisting alkali feldspar andplagioclase in syenite CH2 (dots) and monzodiorite UCV21(squares). Isotherms calculated from Green & Usdansky(1986).

FIG. 2. Hand specimen of a melilite vein (BV) from UCV,sequence W, cross-cutting the wollastonite exoskarn. Thebeige mineral is a single crystal of wollastonite (10 cm inlength). Idiomorphic black titanian garnet occurs betweenwollastonite and melilite (dark grey with white alteration).Yellow vesuvianite appears in the center of the patches ofaltered melilite (white).

1405 39#5-oct-01-2277-15 26/10/01, 13:011410


bution inherited from the intrusive rock have composi-tions in the range 5.0–9.5 wt.% Al2O3 and 0.17–1.24%TiO2 (CH6, CH8). Later-formed very small grains ofdiopside (0.4–3.5 wt.% Al2O3 and less than 1.24% TiO2)occur as scattered aggregates.

The garnet composition is Grs80–94 with less than0.60 wt.% TiO2 (Table 5, Fig. 5). A few aggregates ofgarnet (CH78, CH8) have a size and rectangular shapereminiscent of the idiomorphic plagioclase observed in

the monzodiorite (but not in the “syenite”). They aremuch more scattered and sporadic than the plagioclaseof the main part of the intrusion, and may correspond tophenocrysts. In one case, a blurred zonation with a con-centric idiomorphic outline was observed in this type ofpseudomorph, which may be interpreted as being de-rived from a more sodic zone within this magmatic pla-gioclase.

zone A1, where K- feldspar is replaced by very fine fibrouswollastonite and grossular. In the center, a crystal of titaniteis partly replaced by perovskite. Width of field is 2.6 mm.Plane light. B. Development of wollastonite at the expenseof clinopyroxene along its rims (CH4, boundary A1–A3).Width of field is 2.6 mm. Cross-polarized light. C. Recrys-tallization veins of wollastonite and grossular in the fibrouswollastonite – grossular association of zone A3. Width offield is 2.6 mm. Cross-polarized light. D. Wollastonitearrays in zone A3 close to the exoskarn (sample CH8).Comparatively large crystals of wollastonite are sur-rounded by smaller slender prisms in a fan-like pattern. Theonly other mineral is grossular (isotropic). Width of field is2.6 mm. Cross-polarized light. E. Remnants of the earlypyroxene-bearing endoskarn A close to the boundary withzone B (sample UCVZ). The light- colored rounded grainsare aluminian diopside, the dark ones are grossular, allincluded in a large crystal of wollastonite. Width of field is0.65 mm. Cross-polarized light.

FIG. 4. Photomicrographs of zone A of the endoskarn. A.Boundary between the syenite on the right-hand side (K-feldspar, green clinopyroxene, titanite) and the endoskarn,

1405 39#5-oct-01-2277-15 26/10/01, 13:011411


Zone A2 may be almost indistinct, as in sample CH4,in which the green pyroxene inherited from the intru-sive rock shows a transformation into wollastoniteaggregates progressing along its margins (Fig. 4B).Diopside is only very sporadically observed in thissample, invariably as tiny grains.

Zone A3

Wollastonite and isotropic garnet are the predomi-nant phases. Near the contact with zone A2, the fine fi-brous, parallel texture of wollastonite typical of zoneA2 is still present, and the garnet (Grs95) is fine-grained,except for sporadic recrystallized patches (Fig. 4C).Further from the A2–A3 contact, recrystallized areaspredominate, and wollastonite occurs as coarser more-or-less parallel prisms, but also as large xenomorphicpatches, associated with grains of garnet of similar sizeand shape. The garnet composition is Grs83 with 0.7–0.8 wt.% TiO2. No perovskite could be found either inthis zone or in A2.

In sample CH8, in the part of zone A3 nearest(5 mm) the exoskarn, a definite change is observed inthe original texture of the rock: the slender prisms ofwollastonite which, in this region, play the role of the

fibers of zone A1, are no longer parallel, but rather or-ganized in fan-like pattern (Fig. 4D). The fan hinges areinvariably oriented toward the marble. It is interestingto observe this texture in zone A (but near the exoskarn,in this sample and in CH22), because it is the one wefind (with minor variants) in zone B.

In UCV 60 and 61, where the wollastonite – grossu-lar association does not show any parallel orientation, ayellow titanian garnet (Table 5, Fig. 5) occurs in patches,associated with coarsely recrystallized wollastonite.

The boundary A–B

This boundary is rarely observed. One exception issample CH8, in which zone A3 is in contact with zoneB (somewhat modified). Another one is the wollasto-nite – garnet – vesuvianite association of sample UCVZ,in which wollastonite occurs as fan-like aggregates ofslender prisms, associated with subordinate irregularsmall grains of colorless garnet (Grs50). The garnet andvesuvianite crystals include a multitude of very smallghost crystals, commonly with a distinct rectangularoutline, similar to what is commonly observed to remainof former clinopyroxene crystals in many garnet skarns.In a recrystallization vein, a large crystal of wollasto-

1405 39#5-oct-01-2277-15 26/10/01, 13:011412


nite was found to contain numerous minute rounded (butneither altered nor corroded) inclusions (Fig. 4E), prob-ably preserved as armored relics, among which pyrox-ene is predominant. This pyroxene contains 49 to 54mol.% of the Ca(Al,Fe)2SiO6 end-member and alsocomparatively high Ti contents (e.g., #10/23, Table 4).At a distance of 2 cm, one small grain of melilite (#10/17, Table 6) and idiomorphic grains of grossular (Grs76,0.84 wt.% TiO2) were found as inclusions in other crys-tals of wollastonite. The presence of relics of melilite,aluminian diopside and grossular in this rock suggeststhat it contained, prior to the development of the asso-ciation vesuvianite – wollastonite – garnet, the boundarybetween the zone B and a zone resembling A2, althoughthe pyroxene has a peculiar, Al-rich composition.

Zone B

Zone B is composed of wollastonite and melilite incomparable amounts and in apparent equilibrium. In thiszone and in the coarse-grained melilite vein describedbelow, the composition of melilite shows little varia-tion: 57–65 mol.% Gh with XMg in the range 0.82–0.89(Table 6, Fig. 6). At UCV, it is more Na-rich (4.3 to6 mol.% “Na-melilite”) than at CH (1.8 to 3.1 mol.%“Na-melilite”).

In sample UCV8, melilite occurs mostly as rathersmall crystals of almost uniform size with a certain idi-omorphic tendency, without any observable zoning.Wollastonite occurs as short prisms, of the same size asthe associated crystals of melilite, commonly with amore-or-less parallel orientation. However, from placeto place, melilite shows a striking tendency to an orga-nization in parallel grains that tend to coalesce into

porphyroblasts with a diameter at least ten times largerthan the individual grains (Fig. 7A). These randomlydistributed porphyroblasts constitute an important frac-tion of the rock, and contain numerous very smallrounded crystals of wollastonite.

At CH, a mosaic association of melilite and wollas-tonite is observed, with a weak but distinct orientationof the wollastonite prisms as in UCV8. However, wol-lastonite also shows another type of textural develop-ment, reminiscent of that found in zone A3 (CH8) orA2 (CH22): radiating circular arrays of relatively shortslender prisms, organized around large patches formedof one or a few crystals with a good development offaces (h0l). Melilite tends to segregate or to constituteshort blind veins, and the mosaic texture is interruptedby a network of veins of very coarsely crystallized wol-lastonite ± melilite (Fig. 7B). In CH8, the narrow (lessthan 5 mm) zone B, in contact with the fibrous garnet –wollastonite association of zone A3, consists of pre-dominant coarse melilite (entirely replaced by an ag-gregate of Al-rich vesuvianite) associated with verycoarse, stout prisms of wollastonite.

The wollastonite near the exoskarn boundary is re-sorbed owing to the corrosion of the large patches, ortheir persistence only as small rounded inclusions inmelilite crystals of uniform size (0.5 mm, CH10, CH12,CH13). This is perhaps also the case in CH49, in whichwollastonite occurs as rare clusters of tiny interlocking

FIG. 5. Compositions of garnet from CH and UCV skarns.

1405 39#5-oct-01-2277-15 26/10/01, 13:011413


grains with no tendency to elongate shapes and no ap-parent twinning and, in some cases, as inclusions inmelilite in the vicinity of the clusters.

The only other mineral observed in these rocks (apartfrom secondary veins) is titanian garnet. Its most spec-tacular mode of occurrence is as large yellow crystals,

FIG. 6. Compositions of melilite of the vein stage (mol. fraction) in skarns from CH andUCV (this study), and other skarns (as given by Deer et al. 1986).

FIG. 7. Photomicrographs of zone B of the endoskarn. A.Coalescence of the melilite crystals (bluish grey) intoporphyroblasts that include wollastonite crystals of smallersize than wollastonite not included in melilite. Width offield is 2.6 mm. Cross-polarized light. B. Wollastonite arraysimilar to that of zone A3 (Fig. 4D). Melilite occurs as veinson the right-hand side. The isotropic mineral around and inthe outer part of the “cauliflower” is titanian garnet. Widthof field is 2.6 mm. Cross-polarized light. C. Roundedidiomorphic crystal of titanian garnet, extended byinterstitial veinlets between melilite crystals (“spines”).Width of field is 2.6 mm. Plane light.

1405 39#5-oct-01-2277-15 26/10/01, 13:011414


either as poikilitic patches similar to those mentioned inzone A3, developped on sectors of the circular wollas-tonite arrays (Fig. 7B) and in the wollastonite veins, oras rounded grains (Fig. 7C) with inclusions of wollasto-nite and melilite. The garnet composition is in the rangeGrs26–35, with 5.22–8.11 wt.% TiO2 (Table 5, #14/21).

Granules of pyrrhotite (CH3, CH51) or pentlandite(CH3) (or both) are common, but occur in smallamounts; perovskite is rarely observed as tiny octahedra.

From this point on, there are differences betweensequences S and W, which are thus described separately.

Sequence W: vein BV

The contact of zone B (UCV8) with the melilite veinBV is sharp and characterized by textural changes inmelilite and by the fact that wollastonite loses any traceof subparallel orientation (Fig. 8A). Textural changesin the melilite crystals include a larger average grain-size, an extreme variability in grain-size (0.2 mm –

5 cm), and a striking idiomorphic character. In places,but less commonly than in zone B, a tendency towardnear-parallelism and coalescence of smaller crystals isobserved. A pattern of very fine lamellae of magnetitein two or three orientations occurs locally in the core ofthe melilite crystals, with a very regular and uniformdistribution that suggests an exsolution origin, but it isonly exceptionally and locally observed in the rim(Fig. 8B).

In UCV8, wollastonite disappears after a 1-mm-widezone, where it recrystallizes in large poikilitic amoeba-like patches, replaced by calcite or, less commonly,brownish yellow garnet (Grs31, 10.4 wt.% TiO2) up tothe boundary of the wollastonite zone. From place toplace within a few millimeters of this boundary, rem-nants of poikilitic interstitial coarse wollastonite are in-cluded in calcite or yellow garnet. This garnet especiallyoccurs in contact with zone C, either as individual crys-tals (Fig. 8C) or as an almost continuous fringe (UCV65;Fig. 2B).

FIG. 8. Photomicrographs of the melilite veins (BV) at UCV, sequence W. A. Boundary between zones B (right) and BV (left).Melilite crystals (bluish grey) change to a larger size and idiomorphic shape, and wollastonite (yellow and light-colored)becomes poikilitic, then disappears. Width of field is 2.6 mm. Cross-polarized light. B. Magnetite lamellae in the core ofweakly zoned crystals of melilite. Width of field is 1.2 mm. Cross-polarized light. C. Interstitial brownish yellow titaniangarnet, enclosing idiomorphic crystals of melilite, and extended by paler veinlets (“spines”). The greyish dotted patches inmelilite are described as the granular modification. Width of field is 2.6 mm. Plane light. D. Veinlet of calcite andhydroxylellestadite (light grey) between melilite and wollastonite and cross-cutting wollastonite. Width of field is 2.6 mm.Cross-polarized light.

1405 39#5-oct-01-2277-15 26/10/01, 13:011415


Veinlets of P-rich hydroxylellestadite associatedwith calcite have been observed along the boundary BV– C with a short cross-cutting extension in the wollasto-nite (Fig. 8D). In UCV65, the texture suggests that theveinlets formed before the titanian garnet.

Sequence W: zone C

Any relation with the endoskarn texture is lacking.Zone C is characterized by very large (more than 2 mm,commonly several centimeters) idiomorphic prisms ofwollastonite that mould the adjacent crystals of meliliteof vein BV, but do not penetrate in the intergranularspaces between the melilite crystals, invariably filledwith calcite. Wollastonite is partly corroded by calcitepatches along its cleavages, especially near the bound-ary BV – C. Further from this boundary, calcite is ob-served in the interstices of the wollastonite framework.The titanian garnet is also observed in the zone C nearthe boundary BV – C (Table 5, #14/38), partly replac-ing wollastonite and the associated corroding calcite. Inall cases, the crystals are zoned, and their titanium andiron contents decrease from center to rim (e.g., UCV7,Grs44–49 with 6.7–5.0 wt.% TiO2; UCV6, Grs30–42 with9.3–5.3% TiO2).

Sequence S: vein BV

At CH, the vein composed of melilite with a uni-form grain-size (0.5–1.5 mm) is separated from themuch finer-grained zone B by a regular planar surface,and it may be a few millimeters thick (CH30, CH8; Fig.9A), or constitute the whole of a thin section (CH51). Inthis sample, the idiomorphic crystals of melilite show agrowth zonation without appreciable compositionalchange, and the same inclusions of magnetite lamellaeas described in sequence W. They include granules ofperovskite in their rim (rarely) and of hydroxyl-ellestadite along their boundaries. The composition ofhydroxylellestadite shows small variations (Table 7). Itis PO4-poor and has an atomic ratio S:Si less than 1,attributed to the presence of CO3 substituting for SO4.An early grossular-rich garnet (Grs59; Table 5, #3/20)observed as an inclusion in a melilite crystal providesour only information on the composition of the garnetpresent in the rock at the time of melilite crystallization.

Narrow veinlets, strings or groups of small idiomor-phic crystals of melilite, rooted in the main vein, occurin the first few centimeters of zone C. In CH30, themelilite veinlets develop (with the same grain-size as inthe main vein) either alone (Fig. 9B) or in association

FIG. 9. Photomicrographs of the melilite veins (BV) andwollastonite – tilleyite exoskarns at CH, sequence S. A. Themain melilite-bearing vein (BV) is in contact on both sideswith coarse crystals of wollastonite (yellow, orange) typicalof exoskarn. In the lower right corner, boundary betweenthe wollastonite exoskarn and the much finer-grainedwollastonite – melilite endoskarn, zone B. Width of field is2.6 mm. Cross-polarized light. B. Clusters of equantcrystals of melilite moulding large prisms of wollastonitein the exoskarn. Width of field is 2.6 mm. Cross-polarizedlight. C. Idiomorphic yellow crystals of titanian garnetsurrounded by tilleyite and enclosing small idiomorphiccrystals of melilite. Width of field is 2.6 mm. Plane light.

1405 39#5-oct-01-2277-15 26/10/01, 13:011416


with much larger crystals of tilleyite in all possible in-terstitial (calcite-filled early on) spaces between thelarge prisms of wollastonite in the exoskarn, which showrecrystallization along their margins. At a distance of 1cm, the apophyses of the vein on the exoskarn sidechange into veinlets and clusters. In CH8, in the samesituation, melilite constitutes a dense irregular networkof large idiomorphic crystals, replaced by vesuvianite.

The yellow titanian garnet (e.g., CH1, Grs17–25 with11.5–4.88 wt.% TiO2) partly replaces wollastonite andtilleyite. In this case, it has an irregular and ragged out-line. It also occurs as comparatively large rounded crys-tals that include many small idiomorphic crystals ofmelilite and also wollastonite and tilleyite (Fig. 9C). Oneyellow garnet crystal (CH7, Grs24, 10.6 wt.% TiO2) isassociated with a small aggregate of perovskite gran-ules defining an idiomorphic outline, suggesting thatperovskite could have replaced a titanite crystal.

Sequence S: the spurrite – melilite variant of vein BV

Sample CH49 is the only one composed mainly ofthe spurrite – melilite association. Melilite is finergrained and distinctly more Al-rich than usual (67–77mol.% Gh, with XMg in the range 0.87–0.94; Table 6).

On the inner side, this association is separated fromthe wollastonite – melilite association by the associa-tion hydroxylellestadite – melilite, in which hydroxyl-ellestadite (Table 7) occurs as very small crystals,commonly included in the rim of the melilite crystals.

In the outer part of the spurrite – melilite zone,spurrite and melilite seem in equilibrium. Their grainsizes are similar and very irregular. In the absence ofdeformation, large idiomorphic crystals of melilite incontact with spurrite show a texture similar to that de-scribed at the boundary BV – C at UCV. A distinct band-ing results from variations in their proportions, and thisbanding is confirmed by the distribution and clusteringof very small but numerous grains of perovskite invari-ably included in the melilite bands (Fig. 10A).

Along the contact with zone C, there is evidence ofa complicated process of deformation: the spurrite bandsare folded, and the spurrite crystals are fragmented, andenclosed in undeformed tilleyite, neither showing anytendency of annealing. Short veinlets of idiomorphicmelilite develop in the tension cracks (Figs. 10B, C). Inone such veinlet, a small cluster of perovskite grains isincluded in the melilite. The main plane of deformationis followed on the inner side by a shear zone 5 mm thick,which contains part at least of the hydroxylellestadite-bearing assemblage. In this sheared zone, melilite isgranulated and recrystallized into very small (0.1 mm)interlocking crystals, in the fashion of quartzite in low-grade metamorphism. Spurrite is also recrystallized aspoikilitic porphyroblasts (0.5 mm), with a well-devel-oped (001) face, commonly transverse to the shearplane.

Titanian garnet occurs around and at the expense ofthe perovskite granules in the banded part of the sample,and also in the perovskite-free sheared area, with a regu-lar sieve-like pattern.

Sequence S: zones C and D

The massive spurrite zone and the tilleyite zone havebeen described by Istrate et al. (1978). The grain size ofspurrite is variable and locally very large. The spurritecrystals are flattened and show a predominant develop-ment of face (001) and widespread polysynthetic twin-ning in the same direction. The other known twin plane,(101), is rarely observed and not polysynthetic. Insample CH1, coarse-grained tilleyite partly replacesspurrite, of which corroded remnants are preserved. InCH49, tilleyite appears along a calcite vein as a milli-meter-thick fringe between calcite and spurrite (Fig.10D). There is no evidence of spurrite – tilleyite incom-patibility, but where mutual relationships were ob-served, tilleyite was found to form after spurrite andpartly at the expense of this mineral (Fig. 10E).

In many samples, the exoskarn includes only a well-defined wollastonite – tilleyite zone, following either thezone B (CH20) or the vein BV (CH30). It contains vein-lets of melilite and titanian garnet and large blades ofwollastonite with a corroded outline. In CH8, stoutprisms of wollastonite form a single row with a clearidiomorphic development toward the wollastonite – cal-cite – melilite association of the outer part of the exoskarn.Tilleyite is either preserved or replaced by a fine-grained secondary aggregate of calcite and wollastonite.The transition to the coarse-grained marble takes placewith an idiomorphic relation of tilleyite toward calcite.

PARAGENETIC DESCRIPTION

OF PYROXENE VEINS IN THE MARBLE

The calcite-free veinlets in sample UCV4–1 mainlyconsist of pyroxene, usually associated with garnet(Grs72–80 with 0.12–0.67 wt.% TiO2, Table 5) or, insome cases, with a little green spinel (Mg0.91–0.86Fe2+

0.09–0.14Al1.98Fe3+0.02O4). Cl-bearing vesuvianite

(#3/4 and 3/5, Table 8) appears as a narrow rim alongsome veinlet boundaries. It tends to change to garnet.

Calcite-bearing parts of the sample show regularlyscattered grains of pyroxene similar to the ones in thecalcite-free veinlets. Some of these pyroxene grains areenclosed in calcite, with a distribution in the outer partsof calcite crystals, showing that they mostly developedalong the boundaries of calcite grains and that theseboundaries were later displaced by calcite recrystalliza-tion. A few small crystals of spinel (Mg0.84Fe2+

0.16Al1.91Fe3+

0.09O4) and of phlogopite (19–20 wt.% Al2O3, XMg0.93–0.94), 10 �m in diameter, are enclosed in calciteor, more commonly, in the vicinity of the pyroxene.

1405 39#5-oct-01-2277-15 26/10/01, 13:011417


The composition of the pyroxene strongly dependson the nature of its immediate neighbors. In veinlets orin direct contact with Al-bearing minerals (garnet, ve-suvianite, spinel or phlogopite), it is aluminian diopside(13.5 to 19.4 wt.% Al2O3) with a variable Ti content, upto 2.6 wt% TiO2, quite similar to the aluminian diop-side observed as inclusions in the endoskarn UCVZ(Table 4, Fig. 11). Where in direct contact only withcalcite, the pyroxene has a variable but lower contentAl content. According to the inferred structural formula,all iron is trivalent, and the ratio Fe3+:Al (0.31), approxi-mately constant, corresponds to almost equal mole frac-tions (up to 28%) of esseneite and CaTs components.This constant ratio is interpreted as due to the occur-rence of pyroxene in equilibrium with a garnet of al-most constant composition (Grs72–80), owing to thereciprocal equilibrium between CaTs, esseneite, gros-sular and andradite (Table 9, eq. 46).

PARAGENETIC DESCRIPTION

OF SECONDARY MODIFICATIONS

The monticellite-bearing granularmodification of melilite

A granular type of modification of melilite occurs asrounded patches, commonly with sharp outlines, alongalmost indistinct fractures in the zone BV (Figs. 8C,12A). A veinlet of this type has been observed to crossthe recrystallization boundary (CH8) in only in one case.Near the contact of the zone BV with the wollastoniteexoskarns at UCV, the patches are larger and coales-cent, and the portions of melilite in contact with thewollastonite (and calcite) exoskarns that persist withoutmodification of this type are scanty. Toward the insideof zone BV, this outer part extends as delta-shapedzones of more localized alteration in continuity with theveinlets. At CH, smaller patches of a granular modifi-cation are observed in the melilite veinlets developed inthe exoskarns.

The patches consist of melilite with the same orien-tation as the normal melilite outside, but having a higherbirefringence; this melilite is associated with granulesof monticellite, variable amounts of garnet, and somephases whose high birefringence and compositions in-dicate the CSH family. In one case, a mineral ofrankinite composition (UCV8) was observed, but wol-lastonite was not found. Where such a veinlet crossesthe amoeba-like patches of wollastonite, the wollasto-nite is mostly unaffected. Both melilite and garnet showprogressive shifts toward Al-rich compositions, withincreasing development of the modification, up to Gh92and Grs88, respectively (UCV9). The granular areasshow in some places a further evolution, with roundedcenters of vesuvianite separated from the other miner-als mentioned by a narrow rim of brownish, poorly crys-tallized phyllosilicate-like material.

In sample CH51, a modification of the same type isdeveloped as very narrow regions along a few fractures.The birefringence of melilite progressively increases,and small irregularly scattered granules of monticelliteappear. These veinlets cross pre-existing spines of gar-net (see below). At a few places along these veinlets,rounded spots of the more usual granular patches ap-pear. One of them shows a center of grossular with arim of melilite Gh80–90 80 �m wide and a few grains ofcuspidine. Monticellite is observed at a distance of 0.2mm. Monticellite alone, without compositional changeof the adjacent melilite, occurs as independent veinlets(Fig. 12B). Magnetite and pyrrhotite (in some cases Ni-rich, tending to a pentlandite composition) occursporadically in the same intergranular spaces asmonticellite.

Garnet spines and veins, and vesuvianitein melilite-bearing rocks

Garnet regularly occurs as localized and limited net-works of very thin discontinuous veinlets, called“spines” by Burnham (1959), which appear betweenmelilite grains, and occasionally between melilite and

1405 39#5-oct-01-2277-15 26/10/01, 13:011418


wollastonite or tilleyite (but never spurrite), not relatedto the granular areas but occasionally enclosing them(Fig. 8C). In zone B of UCV8, they occur everywhere(Grs54, with 1.33 wt.% TiO2). In CH5 and CH57, partof the spines occur as crowns around the crystals ofyellow garnet (Fig. 7D), with a regular compositionalchange from the centers of the garnet patches to the outerlimit of the spines. They are also common in zone B ofsequence S. In vein BV of UCV8, a similar garnet(Grs52–55, 0.2–1.85 wt.% TiO2) develops in interstitialspaces between melilite crystals, in an intermediate po-sition between wollastonite and calcite or yellow

titanian garnet, forming a zone 3 mm wide parallel tothe boundary B – C.

Closest to the spines, the magnetite lamellae in-cluded in the core of melilite crystals (UCV8, CH51)are commonly partially replaced by Ti-free garnet(Fig. 12B). Rounded granules of the same garnet(Grs42–65) are commonly included in recrystallized wol-lastonite (UCV8).

In continuity with the spines, thin calcite–garnetveinlets in exoskarns (CH49, spurrite zone) are related

FIG. 10. Photomicrographs of the melilite – spurrite variantof zones BV – C – D at CH, sequence S (sample CH49). A.Alternating bands of melilite (grey), spurrite (brightlycolored) and perovskite (black aggregates within melilitebands). Note the folding in the lower part of the picture.Width of field is 2.6 mm. Cross-polarized light. B. Verylarge broken crystals of spurrite, with veinlets of melilite inthe tension cracks. Width of field is 2.6 mm. Cross-polarized light. C. Sheared spurrite (brown) in contact withidiomorphic melilite (altered, black). Undeformed twinnedtilleyite (pink, green) develops at the expense of spurrite.Width of field is 2.6 mm. Cross-polarized light. D. Coarse-grained calcite (vein from lower left to center) separatedfrom spurrite (grey) by a fringe of tilleyite (brightlycolored). Width of field is 3.5 mm. Cross-polarized light.E. Elongate prisms of spurrite (grey) corroded and partlyreplaced by tilleyite (brightly colored). Width of field is2.6 mm. Cross-polarized light.

1405 39#5-oct-01-2277-15 26/10/01, 13:011419


to the development of tilleyite veins that penetrate thespurrite zone and may follow for some distance thespurrite – melilite boundary. Garnet in these veinlets hasthe same range of composition as the spines (Grs42, 0.01wt% TiO2; Grs52, 0.7 wt.% TiO2). In the same sample,a vesuvianite – calcite symplectite also appears at thetip of a garnet veinlet that branches out into a recrystal-lization veinlet in spurrite; tilleyite – calcite symplectitesand interstitial grains of calcite occur nearest thespurrite. In wollastonite-dominant exoskarns (CH7,CH1), a narrow rim of the same pale yellow garnet(Grs54–72, 2.38–1.43 wt.% TiO2) is observed betweenthe earlier titanian garnet or the melilite crystals andtilleyite, with apophyses between the tilleyite crystals.It is associated in some cases with vesuvianite. Vesuvi-anite crystals with an “idiomorphic” zonation occur incontact with calcite, not far from zone B (Table 8,UCV15, core #15/44, rim #15/42), and may partiallyreplace earlier yellow garnet, as shown by the locallyhigh Ti contents (e.g., #15/8, UCV7).

The boundary BV – C, where modified (UCV9), alsoshows the development of a vesuvianite – garnet asso-

FIG. 11. Compositions of aluminian diopside from endo-skarns and skarn veins in the marble at UCV: Al and Al +Fetotal (in apfu) versus mol. fraction Ca(Al,Fe3+)2SiO6.

1405 39#5-oct-01-2277-15 26/10/01, 13:011420


ciation partly at the expense of the melilite crystals al-ready affected by the granular modification. On thewollastonite side, a yellowish garnet (Grs57, 2.0 wt.%TiO2) forms a rather regular millimetric zone and, inplaces, a network of anastomosed veinlets enclosingislands of wollastonite; these veinlets may connect thismillimetric zone to the rounded crystals of earlierbrownish yellow titanian garnet developed at a shortdistance in the exoskarn. On the melilite side, vesuvian-ite forms oval patches, changing in turns to colorlessgrossular (Grs85–90), including in some places patchesof hibschite (Ca3Al2Si1.7H5.2O12). In UCV65, the sametransformation is widely developed in places where themelilite – wollastonite boundary is not rimmed bytitanian garnet, as shown in Figure 2. In CH30, whichcontains melilite, wollastonite and altered tilleyite, thegrossular + vesuvianite association is developed, partlyat least, at the expense of wollastonite.

In CH49, large roundish patches of vesuvianite areobserved in direct contact with fresh melilite, whichshows over a distance of 15 �m a strong but progres-sive compositional change toward the gehlenite end-member (up to Gh82). The associated garnet is Grs53with 1.27 wt.% TiO2. The composition of vesuvianiteshows small variations with respect to that given inTable 8 (#4/15).

In some of these occurrences, vesuvianite seems tohave been overgrown by (and sealed in) the later yel-lowish garnet.

Recrystallizations and vesuvianite in endoskarns A

Coarsely recrystallized areas tend to develop asbands parallel to the boundary with the intrusive rock.They contain remarkable instances of successivegrowth-zones in the garnet and, in some cases, calcitepatches and even quartz, which are absent in the finer-grained parts. This recrystallization produces large idi-omorphic crystals of grossular, which tend to segregateas irregular veinlets surrounded by large crystals ofwollastonite. Xonotlite is observed occurring side byside with wollastonite (CH 78). A few small prisms ofapatite, included in the garnet, show an incipient substi-tution of P by Si and S. Hydroxylellestadite fills mostof the interstitial spaces between the idiomorphic crys-tals of garnet. Other interstitial spaces in the same areasare filled with clinochlore or “pennine”, or a mixture of“pennine” and chrysotile. Chrysotile was also observedin the same situation associated with minor calcite andrare anhydrite.

Fine-grained subidiomorphic vesuvianite appears aslarge patches, which may predominate (UCVZ) but,more commonly, constitutes only a fraction of the rock(CH4, CH8, UCV60, UCV61); it tends to form discon-tinuous veinlets roughly parallel to the contact with theintrusive rock. Such veinlets locally replace titaniangarnet (UCV 60 and 61). The associated minerals aregrossular, chlorite, chrysotile, calcite and scarce purediopside, in large idiomorphic crystals. In CH6, a cross-

1405 39#5-oct-01-2277-15 26/10/01, 13:011421


cutting vein of diopside, 0.5 mm thick, is surrounded onboth sides by a millimeter-wide zone entirely devoid ofpyroxene, containing some vesuvianite in addition torecrystallized grossular.

The alteration of melilite

In a few samples, melilite is altered along veins toyellow garnet (Grs25–30, 4.07–5.60 wt.% TiO2), withclusters of perovskite (CH3, CH49). The developmentof this garnet starts on earlier-formed grains of titanianyellow interstitial garnet of zone BV. A more-or-lessgradual zonation can be observed from the yellow gar-net to pale yellow (e.g., Grs47, 2.93 wt.% TiO2) and fi-nally colorless garnet (Grs51–67, 2.51–1.27 wt.% TiO2;e.g., #4/8, Table 5) in contact with vesuvianite.

Melilite also may be altered to vesuvianite, abicchulite-like mineral and a kamaishilite-like mineral;small patches of vesuvianite (without calcite) developon melilite (e.g., CH49, #8/17, Table 8). Near such ve-suvianite and even at its contact, the optical continuityof the melilite crystal is preserved (no difference in ei-ther orientation, birefringence or indices of refraction),but the crystal becomes isotropic, with a sharp bound-ary with melilite. The electron-microprobe analyses ofthis isotropic region and its immediate vicinity (Table 6,#8/18) suggest that the cubic mineral is bicchulite, con-sidered by Henmi et al. (1973) as a hydrated melilite-group mineral, and the anisotropic surrounding regionis kamaishilite, which has been interpreted as a tetrago-nal polymorph of bicchulite (Uchida & Iiyama 1981),although the compositions are in fact different.

Alterations of melilite in zone BV also include“bicchulite” occurring as a later alteration along a dis-tinct network of veins. Where this alteration becomesextensive, it is accompanied by the development of ve-suvianite as cores in the same veins. Similar vesuvian-

ite replaces melilite in part of the melilite veinlets inexoskarns (#37/49, Table 8), the replacement of idi-omorphic crystals leads to a peculiar aspect and a com-position richer in Al than usual. Locally, this vesuvianiterecrystallizes into large zoned patches with the usualaspect and composition. In UCV9 and UCV65, largepolycrystalline patches of vesuvianite occur at a fewplaces as pseudomorphs after melilite, side by side withthe predominant granular patches. Neither of them wasobserved to enclose or be modified by the other. On theother hand, vesuvianite has been mentioned to replacethe center of some granular patches.

In zone B, large monocrystalline patches of vesuvi-anite are associated with altered zones in which wollas-tonite is more-or-less preserved, but melilite is alwayscompletely altered (CH3 and CH57). The birefringenceand composition of this vesuvianite (Table 8, #2/10 and2/12) are variable; it has low contents of Ti, F and Clcompared to that described above, but may show a com-plex zonation, with a slight decrease in Al and increasein Fe contents toward the rim.

The alteration of melilite also may lead to a cebollite-like mineral. This product is perhaps the most com-monly observed type of alteration of melilite, andconsists of brownish aggregates of fine fibers, the com-position of which resembles that of cebollite, e.g.,Mg0.53Ca4.04Al2.08Si2.96O12(OH)3.24.

Other alteration assemblages

Other alteration assemblages are mostly restricted towollastonite in veins and remain local. Patches of amineral of foshagite-like composition are observed as aproduct of the incomplete replacement (with clearboundary) of the large crystals of wollastonite in theveins. A mineral of mountainite composition occurs asrare small grains in the same veins. A Cl-bearing vari-

FIG. 12. Photomicrographs of the secondary modifications of the melilite veins. A. Along the boundary between a largeidiomorphic crystal of melilite (partly altered in the lower left part) and a very large prism of wollastonite (exoskarn), calcitereplaces wollastonite, and patches of granular modification (gehlenite – monticellite – grossular) appear in melilite. Width offield is 2.6 mm. Cross-polarized light. B. Veinlets of monticellite (brightly colored) interstitial between melilite crystals thatinclude grains of garnet. Width of field is 2.6 mm. Cross-polarized light.

1405 39#5-oct-01-2277-15 26/10/01, 13:011422


ety is also observed as a late product in the heavily al-tered parts of CH3. Mountainite is a CSH mineral witha Ca:Si ratio lower than foshagite, wollastonite andxonotlite, but containing potassium (in this case, 0.7wt.% K2O). Cl-free djerfisherite, also a K-bearing min-eral, exceptionally occurs as a secondary modificationof pyrrhotite in CH51.

INTERPRETATIONS

One of the problems raised by the high-temperatureskarns at Apuseni is the origin of the massive melilite-dominant rocks. Together with the occurrence oftilleyite, spurrite, monticellite and hydroxylellestadite,their presence is the most spectacular feature of theseskarns. Similar associations of minerals have been ob-served in the monticellite zone of the skarns atCrestmore, California (Burnham 1959), where theywere interpreted to have developed at the expense ofthe Mg-bearing marble. However, there is no Mg-bear-ing marble in the vicinity of CH and UCV, and meliliteat Crestmore never forms monomineralic bodies as inthe Apuseni Mountains, although its composition (Gh60)is, with a few exceptions, in the same range in both oc-currences.

The texture of wollastonite in endoskarns

The parallel-fiber texture of wollastonite (zones Aand B) can be understood as resulting from its growth ata metasomatic front as defined by Korzhinskii (1970),i.e., from the fact that the fluid–rock disequilibrium,therefore chemical reaction, was restricted in space tothe surface that separated the wollastonite zone from theadjacent one.

The usual acicular shape of wollastonite crystals in-dicates that the growth kinetics of the (h0l) faces aremuch slower than the others under conditions of highdisequilibrium. Therefore, the wollastonite crystals thathad their c axis perpendicular to the front grew fasterthan the others, which thus remained within the wollas-tonite zone, at the rear of the front where the fluid–rockinteraction took place, and ceased to grow. A state ofparallel growth was quickly reached, and persisted aslong as the growth process went on.

The rock at the immediate contact of zone A, whereobserved, is always a (quartz-free) “syenite”; in thiscase, the mineral that was transformed into wollastonitewas not quartz but alkali feldspar, and grossular wasformed simultaneously. In a case of limited mobiliza-tion of Al and Si, the resulting rock has a uniform pro-portion of grossular and wollastonite in parallel prisms,which is as a first approximation what is observed inzone A.

There are two discordant observations (CH8 andCH22) in the outer part of zone A3, close to the exoskarnboundary: a fan-like organization of wollastonite con-trasts with its tendency to constitute monomineralic

patches (Fig. 4D). The same textural relation has beennoticed between wollastonite and melilite in zone B(Fig. 7B), with circular arrays of wollastonite prisms,the centers of which are occupied by one or a few largecrystals of wollastonite. In agreement with the assumedmode of development of wollastonite, we believe thatthe arrays are reaction rims around quartz grains. As aresult, (1) the melilite – wollastonite association of zoneB derives from the grossular – wollastonite associationof zone A, and the whole region that displays the fi-brous texture of wollastonite, zones A and B, is anendoskarn, formed by the infiltration into the pre-exist-ing intrusive rock (on the scale of a few centimeters) ofCa-rich fluids circulating in the region of the contact.The boundary between the zones B and C thus corre-sponds to the former boundary of the intrusive rock, inaccordance with the striking difference in grain size andtexture of the wollastonite crystals between these zones.(2) The endoskarn at the inner limit of the exoskarn wasnot developed at the expense of the quartz-free “syen-ite” but at the expense of the quartz-bearing monzo-diorite. The presence, in two samples, of a few largerectangular ghosts of zoned plagioclase (which do notexist in the syenite) is consistent with this interpreta-tion. This conclusion is also supported by the fact thatmelilite, which is by far the predominant Al- and Mg-bearing mineral in the entire zone B, has the same Al:Mgratio as the monzodiorite, not the “syenite” (Fig. 13,analyses in Piret 1997). The “syenite”, which has beenshown to have pegmatitic–hydrothermal features, mostprobably developed as a late evolution of the intrusiverock. The portion of zone A that shows the parallelwollastonite texture formed at its expense.

The “melilite vein” stage and its subsequent evolution

The melilite ± wollastonite zones, observed on bothsides of the former intrusive contact, show occasionalvariations in the sequence of zones (the absence ofmelilite in CH22 and the spurrite – melilite associationof CH49 are the most conspicuous examples). Anotherirregular feature is the proportion of the fibrous-wollas-tonite-bearing endoskarn that belongs to zone B; it ispractically absent in CH8 and well developed in othersamples from CH. As a consequence, the developmentof melilite ± wollastonite cannot be interpreted as be-longing, together with other endoskarns, to a single syn-chronous metasomatic sequence. Melilite developmentwas secondary after, and discordant on, pre-existingskarns. These early skarns, which cannot be preciselydefined by observation owing to their reworking, aretentatively characterized below.

The melilite ± wollastonite associationof the vein stage

At CH, the massive melilite vein that developedalong the endoskarn – exoskarn boundary contains no

1405 39#5-oct-01-2277-15 26/10/01, 13:011423


wollastonite, except along its outer side, where the pro-portion of medium-grained crystals of wollastonite in-creases sharply, up to the rather blurred contact with theexoskarn, formerly composed of wollastonite and prob-ably calcite (and later of wollastonite and tilleyite).However, the vein is not in every case strictly located atthe contact with zone B. In places, it is separated fromzone B by 1–2 mm of massive wollastonite with similarmedium grain-size and texture as the exoskarn border-ing the vein; the change of grain size and texture at thelimit of zone B is very sharp (Fig. 9A).

At UCV, a different type of crystallization (or re-crystallization) of the melilite results in a completelyidiomorphic shape of the melilite crystals, which com-monly are very coarse-grained and show extremechanges of dimensions everywhere. Melilite largely pre-dominates over the associated minerals up to the sharpcontact with the wollastonite – calcite exoskarn, whereonly rare veinlets of coarse melilite are observed. Theextreme variability of grain size is typical of pegmatites,and the porphyroblastic tendency of the melilite in theadjacent part of zone B is reminiscent of a vein influ-ence. The relation of zone BV with zone B is, as a firstapproximation, the same as at CH, except for the ab-sence of the outer part of endoskarns (characterized bythe wollastonite arrays). It is the case with sample

UCV8, which shows a large development of “zone” BVat the expense of endoskarns.

The simplest way to understand the relation betweenthe veins and zone B is to assume that the endoskarn Bdeveloped under the influence of the vein-forming fluid,by transformation of a wider or narrower part ofendoskarn zone A (diopside – grossular – wollastonite)to fine-grained melilite Gh60 + wollastonite, with amore-or-less preserved texture.

Another analogy between the veins BV at CH andUCV is that many samples of the UCV suite show astriking tendency for the proportion of wollastonite todecrease owing to its replacement by melilite. This re-placement, as well as the development of melilite veinsalong the exoskarn boundary and within the exoskarn atCH, involve the addition of Al and Mg. An obvioussource for Mg is zone A2, from which pyroxene disap-peared through its transformation into A3. Clearly, Mgwas mobile in the vein-forming process, and even per-fectly mobile in the vein itself, as shown by the fact thatthe melilite in the vein has exactly the same composi-tion as the melilite in zone B (the sole exception occur-ring in sample CH49), in spite of contrasted modes ofdevelopment and association.

Directly related to this zone of recrystallization isthe incidence of titanian garnet, which replaced wollas-

FIG. 13. Projection of the compositions of monzodiorite, “syenite” and skarn minerals onthe CaO–AlO1.5–MgO plane (mol. fractions).

1405 39#5-oct-01-2277-15 26/10/01, 13:011424


tonite (zones B and C) and started to develop in zone Cpractically at the same time as melilite, as it includessmall idiomorphic crystals of melilite and occurs sideby side with coarser-grained melilite. The crystalliza-tion of titanian garnet in a rock originally composed ofpure calcite implies that titanium also was displaced bythe fluids, circulating on both sides of the B–C bound-ary. Perovskite, which appeared prior to titanian garnetand occurs commonly in endoskarns near the contactwith the “syenite” (and, in one case, the exoskarn), isabsent from the largest part of endoskarns. In mostcases, it disappeared not by replacement by other Timinerals (exceptions are locally observed in CH49,CH7, UCV61 and possibly CH3), but by leaching outof titanium, which was redeposited as titanian garnetduring the vein stage.

Tilleyite appeared at CH in equilibrium with meliliteand before the titanian garnet, as shown by the observa-tion of tilleyite inclusions in titanian garnet that devel-oped in equilibrium with melilite. It developedextensively at the expense of calcite and some parts ofthe pre-existing spurrite and wollastonite, forming azone a few dm thick, at least, in contact with the marbleand penetrating along veins in the inner zones. Tilleyiteremained stable in association with vesuvianite beforeultimately being altered, in many samples, to wollasto-nite + calcite.

The high-temperature granular modification(monticellite stage) and the crystallizationof titanian garnet

The usual pattern of occurrence of monticellite is thegranular modification, which immediately postdated thevein stage and has exactly the same areal extent. It ischaracterized by the development of a three-phase as-sociation (gehlenite, monticellite and a Si–Ca mineral,which seems to have been originally rankinite). Com-pared to the earlier melilite – wollastonite stage, theaddition of one mineral indicates that one of the charac-teristic features of this modification is the increase inthe number of inert components. The appearance ofgarnet in the granular patches can be interpreted in asimilar way, by a further increase of the number of inertcomponents.

The change of a component from a perfectly mobileto an inert status in an abstract notion. In concrete terms,“perfectly mobile” means that the chemical potential ofthis component is ruled by the fluid, which implies long-continued flow of a fluid of constant composition, usu-ally resulting in a decrease in the number of associatedphases at equilibrium. An increase in the number ofphases, limited to a local scale, as observed in the caseof the granular modification, may be expected to resultfrom a reduced flow of fluid associated with a changein physical or chemical conditions, such as a drop intemperature or the infiltration of a different fluid in smallamount, with modifications controlled by the access of

this fluid. Therefore, we believe that from this stage on,the large-scale metasomatic activity of the vein stage isover.

Although titanian garnet appeared during the mainmelilite – wollastonite stage, its largest developmentgenerally followed the granular modification. In suchcases, it was probably in equilibrium with the mineralsthat constitute the granular patches, as suggested by itsdevelopment around the patches, as a replacement ofwollastonite. It first developed in a typical sporadicporphyroblastic manner, which indicates a difficulty innucleation; in most cases, the growth started at the ex-pense of wollastonite. Its zonation, never oscillatory,shows that its Ti content was highest at first (and highlyvariable at a given place, in the range 11–4 wt.% TiO2,with a remarkable correlation with the Mg contents,which reach 0.89 wt.% MgO), and decreased in subse-quent stages of growth. This may distinguish it from thetitanian garnet formed in other skarns related to the low-temperature, mineralization-bearing andradite stage. Asfor Mg, an obvious source for Ti is the endoskarn zoneA, from which this element was leached; but in contrastwith Mg, Ti cannot enter the melilite structure, so thatits precipitation required nucleation of garnet, underprobable conditions of oversaturation. Owing to anoverstepping of the reaction, the growth of garnet wasprobably very rapid at first, then slower, more irregular,using preferentially the intergranular spaces betweenother minerals, but without further nucleation (hence the“spine” texture). Its regular compositional change isconsistent with the assumed initial oversaturation of Tiin the fluid, and also with a decrease in the rate of fluidflow.

The granular modification occurred in relation withfracturing, exceptionally clear in CH51, just as thetilleyite – vesuvianite stage in CH49. The somewhatdiscontinuous character of these stages is consistent withthe idea that these transformations depended on the localavailability of fluids along very discrete microfractures.

The mode of occurrence of vesuvianite

When compared to other skarns in which it plays anessential role, vesuvianite is striking at CH and UCV byits late, irregular and sporadic mode of occurrence. Itforms a secondary zone developed by reaction betweenzones A and B, as suggested by the relation observed inUCVZ. Together with grossular, it forms another sec-ondary zone between the melilite crystals, already af-fected by the granular modification, and the wollastonitein zone C. In this case, vesuvianite alternates with thepale yellow garnet developed later than the grossular.Garnet – calcite – vesuvianite veins, associated with thetilleyite veins (CH49), were interpreted as the productof the local transformation of melilite + spurrite. Vesu-vianite also is a product of secondary transformationsin regions with Mg-bearing minerals, without relationto the zonation, for example as a pseudomorph after

1405 39#5-oct-01-2277-15 26/10/01, 13:011425


melilite in the exoskarns and patches in the endoskarns.It may be one of the alteration products of melilite, post-dating the granular modification. Finally, it may evolvefrom the “kamaishilite” – “bicchulite” type of alteration,of which it constitutes the third product.

Vesuvianite occurrence is thus invariably secondary,and restricted to places where Mg was locally available,suggesting that its irregular distribution is due to its lateformation, at a stage at which Mg was no longer mo-bile.

The spurrite – perovskite – melilite associationof the first-stage endoskarns

The occurrence of perovskite pseudomorphs aftertitanite in the inner part of zone A1 and in one exoskarnsample indicates that Si was leached out while Ti had aperfectly inert behavior, prior to the vein stage. The moststriking example of a preserved original distribution ofthe Ti minerals among CH and UCV skarns is providedby the banded sample CH49, which is also the onlyperovskite-rich rock of our collection. The developmentof spurrite + melilite in this sample, instead of the moreusual wollastonite + melilite association, accounts for astronger undersaturation in silica. In another spurrite-bearing sample (CH1), spurrite is observed as relics withrespect to tilleyite. Even in CH49, there are many indi-cations that the spurrite – melilite association is a relictone; on the outer side, the reworking of the spurrite intotilleyite ± calcite, and on the inner side, the modifica-tion of melilite to gehlenite at the contact with grains ofvesuvianite, are indications of the role of the later fluidsresponsible for the high-temperature modifications.

Taking into account also the exceptional narrownessof the zone system in this sample (2 cm) and the mo-saic, metamorphic-like texture of the inner part of thesample, it is probable that only a small amount of fluidwas involved in the (re)crystallization of the observedassociation. The difference in composition of melilitebetween this sample and all others, mentioned above,may be more than merely accidental, and may indicatethat its Mg:Al ratio was not affected by the uniformi-zation in melilite composition so conspicuous else-where. We suggest that this rock was little affected bythe major influx of fluid characterized by Ti and Mgmobility, and that its strong undersaturation in silica wasreached in the early, Ti- and Mg-inert stage.

The grossular – aluminian diopside associationof the first-stage endoskarns

The absence of any Mg-bearing mineral in zone A3has been interpreted as an indication of reworking ofthis zone at the vein stage. Some indications of the ear-lier association of minerals were found in sampleUCVZ; it contains inclusions of grossular and aluminiandiopside very different in both composition and texturefrom those of the intrusive rocks and zone A2. The only

other observed occurrence of the same association is thevein network in marble (UCV4–1), with striking simi-larities in the compositions of minerals (Fig. 11). There-fore, we suggest that they have the same origin. As, inthe case of sample UCVZ, the original material is theintrusive rock, the veins in sample UCV4–1 may havebeen derived from magmatic veins in the marble (theperovskite aggregate replacing titanite, observed inexoskarn sample CH7, also suggests the former pres-ence of an apophysis of the intrusive body in the sur-rounding exoskarn).

Owing to the immediate proximity of a melilite crys-tal (also relict in a wollastonite vein), the grossular –aluminian diopside association in sample UCVZ is in-terpreted as the former zone A of the endoskarn, at theexpense of which zone B developed. In addition, thecrystals of aluminian diopside in sample UCVZ andsome of those in UCV4–1 have high Ti contents, indi-cating that Ti was not leached out of these rocks whenthe pyroxene crystallized. This association is thus inter-preted as characteristic of the outer part of the earlyendoskarn, considered as the first stage.

Early fractures in the first-stage endoskarns

Another possible confirmation of the early develop-ment of these associations is found in the fact thatsamples UCV4–1 and CH49 are the only ones that showevidence of fracturing, and shearing of the endoskarn –exoskarn boundary in the case of CH49. Such a strongdeformation was not observed in the wollastonite crys-tals of CH or UCV, except to a moderate extent inCH30, although deformation of wollastonite is com-monly observed in skarns in general. Our tentative con-clusion is that most of the wollastonite – melilite skarnsdeveloped after this episode of deformation. The irregu-lar nature and discordant character of the melilite “zone”in many samples, stressed above as a strongly distinc-tive feature of the modes of occurrence of melilite, areperhaps due to the existence of pathways for later fluids,opened at the time of fracturing.

PHYSICOCHEMICAL CONSTRAINTS

FROM THERMOCHEMICAL MODELING

Thermochemical calculations

The conditions of stability of the mineral associa-tions described have been calculated as function of tem-perature, pressure of CO2 and activity of aqueous silica(Figs. 14–18, Table 9). The THERMOCALC databaseand software (Holland & Powell 1998) were chosen formost calculations, as they include data for spurrite,tilleyite and vesuvianite. We assumed that the garnet onthe join grossular–andradite and melilite on the joingehlenite–åkermanite are ideal, respectively two-siteand one-site solid solutions (Perchuk & Aranovich1979, Charlu et al. 1981). Although no model for vesu-

1405 39#5-oct-01-2277-15 26/10/01, 13:011426


vianite solid-solution is available, the vesuvianite com-position available in the THERMOCALC database[Ca19Mg2Al11Si18O69(OH)9] was considered closeenough to the measured ones to be used. Monticellitecould not be considered; according to the database,monticellite should not be stable in association with ei-ther gehlenite-rich melilite or rankinite, contrary to ourobservations.

As the compositions of pyroxene encountered havea comparatively high content of the esseneite end-mem-ber, for which no thermochemical data are available inthe THERMOCALC database, the set of equilibria in-volving this mineral (Fig. 18) was calculated from Sack& Ghiorso (1994a, b) for the thermodynamic propertiesof pyroxene, melilite and plagioclase, and from Berman(1988) for other minerals, as recommended by Sack &Ghiorso (1994b). In all calculations, the CaTs andesseneite mole fractions in aluminian diopside are takento be equal, in accordance with the analytical data(Fig. 11).

With the exception of monticellite, the calculatedequilibria were found to be consistent with the observedassociations and compositions of minerals. The consis-tency between calculations from THERMOCALC andBerman + Sack databases was checked for the equilib-rium (4) [Gh]–Wo–Cal–[Grs]–CO2; the fugacity of CO2was calculated to be 50 bars (750°C) in the former caseand 68 bars in the latter. We did not correct for this dis-crepancy.

The activities and molalities of aqueous silica, takento be equal, have been determined from the solubility ofquartz in H2O (Manning 1993).

The pressure of H2O

The only OH-bearing mineral found in the high-tem-perature associations of skarn minerals is vesuvianite

(except for xonotlite, for which no thermodynamic dataare available). It developed in the granular patches ofmelilite after the association of melilite Gh92 with themineral of rankinite composition and garnet Grs88(UCV9), in apparent equilibrium with wollastonite. Inthe system SiO2 – Al2O3 – CaO – MgO – H2O (Fig. 14),the invariant point vesuvianite – melilite (Gh92) – gar-net (Grs88) – wollastonite – rankinite is located at 712°Cand 750 bars H2O pressure; at lower H2O pressure, ve-suvianite (eq. 11, Table 9) appears at a lower tempera-ture than rankinite (eq. 10), in accordance with theobservation. Therefore, H2O pressure was lower than750 bars, in agreement with the approximate value of500 bars estimated for the magmatic stage. The stabil-ity conditions of the mineral associations have beencalculated at 750 bars, which may lead to an overesti-mation of temperatures for the vesuvianite-bearing as-sociations, but this choice has no appreciable bearingon the temperature ranges of associations of OH-freeminerals.

The vein stage (melilite)

The association melilite – wollastonite is unstablebelow a temperature that has a small dependence on CO2pressure: 722°C (low pressure of CO2) to 741°C (wol-lastonite – calcite equilibrium) (Fig. 15). The composi-tion of garnet in equilibrium with melilite gives furtherconstraints on temperature. As it depends not only ontemperature, but also on the activity of silica, as shownby equilibria (38) to (40) in Table 9, the values corre-sponding to various associations of minerals have beencalculated (Fig. 16). The composition chosen for Fig-ure 15 (Grs76) is that of the earliest garnet found in ap-parent equilibrium with melilite and wollastonite atUCV. It indicates 741°C as the lowest temperature forthe melilite-bearing skarns [curve (12), square in Fig.16]. In the hydroxylellestadite-bearing sample CH51,the primary garnet is Grs59. As hydroxylellestadite oc-curs associated with melilite in CH49, with a composi-tion similar to that of CH51, between spurrite – meliliteand wollastonite – melilite “zones”, we assume that thismineral is stable at an activity of silica close to that ofthe wollastonite – spurrite equilibrium. For a pressureof CO2 of 11 bars, determined to be the lowest one con-sistent with the garnet (Grs59) – melilite – wollastonite– spurrite equilibrium, the corresponding temperatureis 754°C [dot on curve (8), Fig. 16]. A temperature of750°C is thus considered to provide a realistic represen-tation of the vein stage.

All the mineral associations described in zones B, Cand D imply a low pressure of CO2. The sequence W ofUCV corresponds to a CO2 pressure between 22 bars at750°C (direct contact between calcite and wollastonite,without spurrite or tilleyite) and 26 bars (garnet Grs76 inequilibrium with melilite and wollastonite).

For the sequence S at Cornet Hill, the stability ofspurrite implies a maximum CO2 pressure of 16 bars,

FIG. 14. P(H2O) – T relationships between melilite (Gh92),garnet (Grs88), vesuvianite, wollastonite and rankinite.Abbreviations and reaction stoichiometries in Table 9.

1405 39#5-oct-01-2277-15 26/10/01, 13:011427


whereas a minimum of 11 bars has been inferred fromthe association in sample CH51. As we have concludedthat some of the spurrite-bearing associations precededthe melilite – wollastonite vein stage, the question ariseswhether the associated fluids were different. Assumingthat the pressure of CO2 remained constant while thetemperature decreased after the end of the main stage offluid flow, the presence of tilleyite still stable at thevesuvianite stage (and generally subsequently altered)would indicate that the pressure of CO2 was less than13 bars (Fig. 15), a value consistent with the stability ofspurrite at 750°C and with the assumption that spurrite,melilite and tilleyite were formed under similar condi-tions of CO2 pressure.

The stability fields of the mineral associations in theskarns have been calculated at 750°C as a function oflog a(SiO2) (Fig. 17). The melilite – wollastonite – py-roxene equilibrium, which separates zones B and A, hasbeen calculated for the pyroxene in equilibrium with

melilite Gh60. The calculated composition (Di76CaTs12Ess12, i.e., 8.5 wt% Al2O3) is in the range found forpyroxene in zone A2 (CH6, CH8). Silica activity is ob-served to decrease throughout the entire sequences fromzone A to calcite. The limit titanite – perovskite, notrepresented, corresponds to a(SiO2) = 10–2.5, a valueconsistent with the position of this front between thepyroxene – wollastonite – garnet endoskarn and the “sy-enite”. In the melilite-bearing zones, increasing activityof silica resulted in increasing grossular content of gar-net, from Grs50 in equilibrium with spurrite – calcite, toGrs65 in equilibrium with spurrite – wollastonite, andGrs81 in equilibrium with melilite – pyroxene, in accor-dance with the observed compositional ranges. Thesilica concentration of the fluid in the melilite-bearingzones was very low, in the range 2.9 to 14.5 � 10–4 m,or 7–35 ppm Si.

The development of tilleyite instead of spurrite, instable association with melilite and commonly wollas-

FIG. 15. Low P(CO2) portion of the T – P(CO2) diagram calculated in the system SiO2–CaO–Al2O3–MgO–H2O–CO2. Theconditions of the vein stage at CH (S, melilite – wollastonite – spurrite – tilleyite) and UCV (W, melilite – wollastonite –calcite) are indicated by the grey areas, those of the high-temperature modifications, by the dashed areas. Reactionstoichiometries are indicated in Table 9.

1405 39#5-oct-01-2277-15 26/10/01, 13:011428


FIG. 16. Garnet composition at equilibrium with melilite, vesuvianite and eitherwollastonite, rankinite, tilleyite or spurrite, for melilite Gh60 (solid lines) and Gh90(dashed lines).

tonite, most probably results from a lower temperature(Fig. 15). Tilleyite is stable at 730°C in a range of fluidcompositions that corresponds at 750°C to spurrite andto the spurrite – calcite boundary (Fig. 17), in agree-ment with the largest development of tilleyite on themarble side of the exoskarn. This interpretation is allthe more likely as the development of tilleyite in equi-librium with melilite is followed by the tilleyite – vesu-vianite association, which is stable only below 730°C.The circulation of the same fluid, mainly along fractures,is probably responsible for the locally observed replace-ment of wollastonite by tilleyite, indicating leaching ofsilica in the inner part of the exoskarns.

Secondary modifications

As indicated above, the association constituting themonticellite-bearing granular patches is inferred to beunstable. However, as the progressive destabilization ofthe åkermanite component of melilite into monticellite+ rankinite or OH-bearing Ca–Si minerals is similar tothe experimentally observed decomposition of åkerma-nite into monticellite + wollastonite below 700°C(Harker & Tuttle 1956), it probably occurred under de-creasing temperature.

The association of vesuvianite with grossular(Grs85–90) and gehlenite in equilibrium with tilleyite(CH49) would seem to correspond to a temperature inthe range 710–720°C at 750 bars (Fig. 16), and to asomewhat lower temperature at lower pressure of H2O,in agreement with that inferred above from the appear-ance of vesuvianite in the granular patches in UCV9(T < 712°C); this temperature was thus considered tohave been achieved after the end of the vein stage.

Pyroxene veins and endoskarns of the first stage

The conditions of stability during the developmentof the network of pyroxene veins in the marble at UCV(Fig. 18) have been calculated for one of the most Al-rich pyroxene compositions (Di48, #4/45, similar to therelict endoskarn pyroxene in UCVZ #10/43), and for atypical pyroxene in contact with calcite (Di70, #1/2).

At 750°C, the garnet – spinel – pyroxene associa-tion of the veins is in equilibrium with calcite (eq. 28)under a CO2 pressure ranging from 95 bars (for the py-roxene in contact with calcite) to 240 bars (for the mostAl-rich one). Melilite would appear instead of pyrox-ene at a lower pressure of CO2, shown for the usualcomposition of melilite (Gh60) and also for the compo-

1405 39#5-oct-01-2277-15 26/10/01, 13:011429


sition (Gh64) at equilibrium with the calcite – spinel –pyroxene (Di70) association. According to the previ-ously proposed interpretation of these veins as originat-ing from apophyses of the monzodiorite, the observedassociation of minerals would correspond to their re-crystallization (and leaching of Si, K, Na, Fe) undercomparatively high pressures of CO2. The plagioclasecompositions represented correspond to the range (An35to An70) observed in the monzodiorite. The reactionbetween calcitic marble, diopside and anorthite compo-nent of plagioclase, for example, is expected to producegarnet and aluminian pyroxene (eq. 26); the pyroxenewould become enriched in Al until feldspar is exhausted(e.g., composition #4/45, Table 4). Further reaction ofpyroxene with calcite would produce spinel and garnetand a more diopside-rich pyroxene (eq. 28; e.g., com-positions #1/18, #1/2).

DISCUSSION AND CONCLUSIONS

The changes in CO2 and SiO2 contentsof the fluid through metasomatic reactions

The pressure of CO2 corresponding to the associa-tions of skarn mineral depends on the initial composi-tion of the fluid and on the exchange of CO2 and SiO2between fluid and rock according to reactions such as,for example, the replacement of calcite by spurrite (eq.

33, Table 9). In a marble composed only of calcite, theamount of CO2 produced by this reaction is limited bythe availability of Si in the fluid. As the fluid in equilib-rium with spurrite at 750°C, 750 bars has a maximumconcentration of silica of 2.9 � 10–4 m, the maximumchange in the CO2 concentration in the fluid due to thereplacement of calcite by spurrite, corresponding to theexhaustion of aqueous silica, is 8.7 � 10–4 mole/kgH2O, or, when expressed as pressure, 10–2 bars. Even inthe case of a quartz-saturated fluid (0.046 m), with apressure of CO2 high enough to allow for a direct re-placement of calcite by wollastonite, the gain in pres-sure of CO2 would be no more than 0.6 bars. The smallinfluence of skarn-forming reactions on the pressure ofCO2, in constrast with their major effect on the concen-tration of silica in the fluid, implies that the mineral as-sociation that developed under a small to moderateH2O:rock ratio corresponds to the CO2 pressure of theinfiltrating fluid, whereas its SiO2 activity is stronglyinfluenced by the SiO2 content of the original rock, i.e.,CO2 is perfectly mobile whereas SiO2 is inert, accord-ing to the terminology of Korzhinskii (1970).

A more important source of CO2 in exoskarn-form-ing processes may be the dissolution of calcite, whichmostly depends on the HCl content of the fluid. The HClcontent of the high-temperature, low-pressure fluid isexpected to be high owing to the general tendency ofratios such as HCl/NaCl and HCl/KCl to increase with

FIG. 17. Stability fields of the minerals from sequences S and W as function of pressure ofCO2 and activity of aqueous SiO2, calculated in the system SiO2–CaO–Al2O3–MgO–H2O–CO2 at 750°C (solid lines) and 730°C (dashed lines). The approximate locationsof the zones A to D are indicated for each sequence. Abbreviations and reactionstoichiometries are presented in Table 9.

1405 39#5-oct-01-2277-15 26/10/01, 13:021430


increasing temperature and with decreasing pressure,where the fluid is buffered by a mineral association, suchas is the case in those unmixing from a crystallizingmagma (Shade 1974, Montoya & Hemley 1975, Popp& Frantz 1980, Sverjensky et al. 1991). This tendency,displayed by HCl-rich volcanic gases, has been investi-gated by Shinohara & Fujimoto (1994) at 600°C in thesystem albite – andalusite – quartz – H2O – HCl. At 400bars, the HCl molality of the vapor phase reaches 0.5 m,with a molality of NaCl of 0.3 m. Compared to the sys-tem H2O–NaCl, the addition of HCl strongly increasesthe concentration of Cl that can be reached by the H2O-rich fluid phase at low pressure without unmixing abrine (which would contain 75 wt.% NaCl in this case).With the assumption that the skarns of the vein stageformed in response to aqueous fluids of magmatic ori-gin, these were probably in large part HCl-rich and ableto dissolve calcite. An initial HCl content of 0.5 m wouldresult in the dissolution of 0.25 moles CaCO3 per kgH2O, and an increase in pressure of CO2 of 3.4 bars(Appendix).

Therefore, the changes in the pressure of CO2through fluid – marble interaction are expected to besmall and mainly controlled by the initial HCl contentof the fluid. As lower-pressure conditions result in ahigher HCl content, the difference between the two se-quences might be partly related to the difference in H2O

pressure at the magmatic stage, found to be lower atUCV than at CH.

The expected major changes in the silica content ofthe fluid are consistent with the observed sequences ofzones. In the well-defined range of CO2 pressure of se-quence W, the melilite – wollastonite association isstable under a narrow range of silica concentration, be-low which the association is melilite – calcite. Accord-ingly, wollastonite is replaced by calcite within vein BV,and calcite – wollastonite is the usual association of theexoskarn at UCV. In a similar way, the development ofwollastonite instead of spurrite or tilleyite at theendoskarn – exoskarn boundary in many samples at CHcan be accounted for by an only slightly higher concen-tration of silica, such as that resulting from a higherH2O:rock ratio. In contrast, owing to the small depen-dence of the calcite – spurrite boundary on SiO2 activ-ity (Fig. 17), for a pressure of CO2 in the range inferredat CH, spurrite (or tilleyite) is stable instead of calciteeven under an extremely low concentration of silica. Inaccordance, the marble was widely replaced by spurriteand tilleyite in sequence S.

Origin of melilite, spurrite and tilleyite

The striking difference in pressure of CO2 betweenthe aluminian diopside veins and the melilite-bearing

FIG. 18. Stability conditions of the aluminian-diopside-bearing associations of the first-stage endoskarns (veins in the marbleand inclusions in endoskarn UCVZ), calculated in the system SiO2–CaO–Al2O3–MgO–H2O–CO2.

1405 39#5-oct-01-2277-15 26/10/01, 13:021431


associations implies that the CO2-poor fluid responsiblefor the melilite stage circulated only in very smallamounts in the region of the sample UCV4–1. Recrys-tallization of the pyroxene-bearing associations has beenobserved, with a large range of pyroxene compositions,which indicates a lack of equilibration, in accordancewith the previously stated hypothesis and with the as-sumption that the pyroxene veins represent remnants ofan early association developed at the margins of the in-trusive rock (and along its apophyses) by a CO2-richfluid, as the first-stage endoskarns. Owing also to thetendency toward silica depletion of the fluid associatedwith this first stage, as proved by the crystallization ofperovskite and of silica-deficient pyroxene, the sur-rounding marble may be assumed to have played therole of local source for this fluid, the origin of whichremains unconstrained.

The spurrite – melilite – perovskite association inCH49 has already been interpreted as also indicative ofa limited influence of the vein-stage fluid, although boththe presence of melilite and the low pressure of CO2implied by spurrite are typical of this stage. One of itspeculiarities is the higher undersaturation of this asso-ciation in silica with respect to the melilite – wollasto-nite association in the veins and endoskarns. At theimmediate contact of these veins, the exoskarn consistsof wollastonite, which implies an input of Si on themarble side of the vein system. Even more silica-rich isthe melilite-free zonation of CH22, in which the wol-lastonite exoskarn is in contact with zone A3, and wherethe titanian garnet and the evidence of Mg leaching arewitnesses of the influence of the vein-forming fluid. Asthe silica undersaturation of the first stage is expectedto persist to some extent through the vein stage, varia-tions in silica activity of the associations developed bythe vein fluid along the endoskarn – exoskarn boundarymay result from an irregular development of either theprimary skarns or the vein-stage skarns. Taking intoaccount the comparatively high solubility of silica inhigh-temperature aqueous fluids, the persistence of astrong depletion in silica over a large part of theendoskarns indicates that the overall amount of fluidinvolved in the vein stage was not very large.

As a conclusion, we believe that melilite as well asspurrite and tilleyite are the result of the circulation ofthe vein-stage fluid at the boundary between the marbleand the earlier-formed grossular – aluminian diopsideendoskarns, strongly depleted in Si (and also Fe, Na andK) compared to the original monzodiorite. Many linesof evidence suggest that this fluid was exsolved fromlower parts of the intrusive body at the pegmatite-form-ing stage: (1) its capacity to transport elements usuallycharacterized by a low mobility, (2) its temperaturerange, similar to that found for the end of the two-feld-spar crystallization in the monzodiorite, (3) the pegma-tite-like texture observed at UCV, and (4) the depositionat the end of this stage of P-, S- and F- bearing minerals(ellestadite, cuspidine).

ACKNOWLEDGEMENTS

We are grateful to the Geological Institute of Roma-nia for providing logistical support during the fieldwork. Thanks are due to I. Katona for providing addi-tional samples, organizing field work and taking photo-graphs, to H. Remy and C. Richard for technicalassistance in performing the numerous electron-microbrobe analyses, to S. Baudesson, G. Badin and G.Drouet for preparing the thin sections, to M.N. Sarciafor drawing the figures, and to P. Sonnet for helpful dis-cussions. We are especially indebted to Danielle Velde,whose interest and advice throughout this study havebeen greatly appreciated. We also thank B. Moine andan anonymous reviewer for interesting and fruitful com-ments on a first version of this paper.

REFERENCES

BERMAN, R.G. (1988): Internally-consistent thermodynamicdata for minerals in the system Na2O–K2O–CaO–MgO–FeO–Fe2O3–Al2O3–SiO2–TiO2–H2O–CO2. J. Petrol. 29,445-522.

BURNHAM, C.W. (1959): Contact metamorphism of magnesianlimestones at Crestmore, California. Geol. Soc. Am., Bull.57, 879-928.

________ (1979): Magmas and hydrothermal fluids. InGeochemistry of Hydrothermal Ore Deposits (2nd edition,H.L. Barnes, ed.). John Wiley and Sons, New York, N.Y.(71-136).

CHARLU, T.V., NEWTON, R.C. & KLEPPA, O.J. (1981):Thermochemistry of synthetic Ca2Al2SiO7 (gehlenite) –Ca2MgSi2O7 (åkermanite) melilites. Geochim. Cosmochim.Acta 45, 1609-1617.

DEER, W.A., HOWIE, R.A. & ZUSSMAN, J. (1986): Rock-Forming Minerals. 5. Non-Silicates (2nd edition).Longman, London, U.K.

GREEN, N.L. & USDANSKY, S.I. (1986): Ternary-feldsparmixing relations and thermobarometry. Am. Mineral. 71,1100-1108.

GROSS, S. (1977): The mineralogy of the Hatrurim Formation,Israel. Bull. Geol. Survey Israel 70.

GUNTER, W.D. & EUGSTER, H.P. (1978): Wollastonitesolubility and free energy of supercritical aqueous CaCl2.Contrib. Mineral. Petrol. 66, 271-281.

GUY, B., SHEPPARD, S.M.F., FOUILLAC, A.M., LE GUYADER, R.,TOULHOAT, P. & FONTEILLES, M. (1988): Geochemical andisotope (H,C,O,S) studies of barren and tungsten-bearingskarns of the French Pyrenees. In Mineral Deposits withinthe European Community (J. Boissonas & P. Omenetto,eds.). Springer, Berlin, Germany (53-75).

HARKER, R.I. & TUTTLE, O.F. (1956): The lower limit of stabilityof åkermanite (Ca2MgSi2O7). Am. J. Sci. 254, 468-478.

1405 39#5-oct-01-2277-15 26/10/01, 13:021432


HENMI, C., KUSACHI, I., HENMI, K., SABINE, P.A. & YOUNG,B.R. (1973): A new mineral bicchulite, the natural analogueof gehlenite hydrate, from Fuka, Okayama Prefecture,Japan and Carneal, County Antrim, Northern Ireland.Mineral. J. 7, 243-251.

HENMI, K., KUSACHI, I. & NUMANO, T. (1971): Contactminerals from Kushiro, Hiroshima Prefecture. 1. Gehleniteand hydrogrossular. J. Mineral. Soc. Japan 10, 160-169.

HOLLAND, T.J.B. & POWELL, R. (1998): An enlarged andupdated internally consistent thermodynamic dataset withuncertainties and correlations: the system K2O–Na2O–CaO–MgO–MnO–FeO–Fe2O3–Al2O3–TiO2–SiO2–C–H2O–O2. J. Metamorph. Geol. 8, 89-124.

HUGHES, C.J. (1960): An occurrence of tilleyite-bearinglimestones in the Isle of Rhum, Inner Hebrides. Geol. Mag.97, 384-388.

ISTRATE, G., STEFAN, A. & MEDESAN, A. (1978): Spurrite andtilleyite in the Cornet Hill, Apuseni Mountains, Romania.Rev. Roumaine Géol., Géophys. et Géogr., Géologie 22,143-153.

JAMTVEIT, B., BUCHER-NURMINEN, K. & STIJFHOORN, D.E.(1992): Contact metamorphism of layered shale–carbonatesequences in the Oslo Rift. I. Buffering, infiltration and themechanisms of mass transport. J. Petrol. 33, 377-422.

________, DAHLGREN, S. & AUSTRHEIM, H. (1997): High-gradecontact metamorphism of calcareous rocks from the OsloRift, southern Norway. Am. Mineral. 82, 1241-1254.

JOESTEN, R. (1974): Pseudomorphic replacement of melilite byidocrase in a zoned calc-silicate skarn, Christmas Moun-tains, Big Bend region, Texas. Am. Mineral. 59, 694-699.

________ & FISHER, G. (1988): Kinetics of diffusion-controlledmineral growth in the Christmas Mountains (Texas) contactaureole. Geol. Soc. Am., Bull. 100, 714-732.

KORZHINSKII, D.H. (1970): Theory of Metasomatic Zoning.Clarendon, Oxford, U.K.

MANNING, C.E. (1993): The solubility of quartz in H2O in thelower crust and upper mantle. Geochim. Cosmochim. Acta58, 4831-4839.

MARINCEA, S., BILAL, E., VERKAEREN, J., PASCAL, M.-L. &FONTEILLES, M. (2001): Superposed parageneses in thespurrite-, tilleyite- and gehlenite-bearing skarns from CornetHill, Apuseni Mountains, Romania. Can. Mineral. 39,1435-1453.

MONTOYA, J.W. & HEMLEY, J.J. (1975): Activity relations andstabilities in alkali feldspar and mica alteration reactions.Econ. Geol. 70, 577-583.

NOCKOLDS, S.R. (1947): On tilleyite and its associated mineralsfrom Carlingford, Ireland. Mineral. Mag. 28, 151-158.

PERCHUK, L.L. & ARANOVICH, L.Y. (1979): Thermodynamicsof minerals of variable composition: andradite–grossularite

and pistacite–clinozoisite solid solutions. Phys. Chem.Minerals 5, 1-14.

PIRET, R. (1997): Minéralogie et géochimie des skarns de hautetempérature des régions de Magureaua Vatei et de CornetHill (Monts Apuseni). Mémoire de Licence, UniversitéCatholique de Louvain, Belgique.

POPP, R.K. & FRANTZ, J.D. (1980): Mineral–solution equilibria.III. The system Na2O–Al2O3–SiO2–H2O–HCl. Geochim.Cosmochim. Acta 44, 1029-1037.

REVERDATTO, V.V. (1970): Pyrometamorphism of limestonesand the temperature of basaltic magmas. Lithos 3, 135-143.

________, PERTSEV, N.N. & KOROLYUK, V.N. (1979): PCO2–T-evolution of zoning in melilite during the regressive stageof contact metamorphism in carbonate-bearing rocks.Contrib. Mineral. Petrol. 70, 203-208.

SABINE, P.A. (1975): Metamorphism processes at hightemperature and low pressure: the petrogenesis of themetasomatized and assimilated rocks of Carneal, Co.Antrim. Phil. Trans. R. Soc. A280, 225-269.

SACK, R.O. & GHIORSO, M.S. (1994a): Thermodynamics ofmulticomponent pyroxenes. I. Formulation of a generalmodel. Contrib. Mineral. Petrol. 116, 277-286.

________ & ________ (1994b): Thermodynamics of multi-component pyroxenes. III. Calibration of Fe2+(Mg)–1,TiAl2(MgSi2)–1, TiFe3+

2(MgSi2)–1, AlFe3+(MgSi)–1,NaAl(CaMg)–1, Al2(MgSi)–1 and Ca(Mg)–1 exchangereactions between pyroxenes and silicate melts. Contrib.Mineral. Petrol. 118, 271-296.

SHADE, J.W. (1974): Hydrolysis reactions in the SiO2-excessportion of the system K2O–Al2O3– SiO2–H2O in chloridefluids at magmatic conditions. Econ. Geol. 69, 218-228.

SHINOHARA, H. & FUJIMOTO, K. (1994): Experimental study inthe system albite – andalusite – quartz – NaCl – HCl – H2Oat 600°C and 400 to 2000 bars. Geochim. Cosmochim. Acta58, 4857-4866.

STEFAN, A., ISTRATE, G. & MEDESAN, A. (1978): Gehlenite incalc-skarns from the Magureaua Vatei–Cerboaia (ApuseniMountains – Romania). Rev. Roumaine Géol., Géophys. etGéogr., Géologie 22, 155-160.

SVERJENSKY, D.A., HEMLEY, J.J. & D’ANGELO, W.M. (1991):Thermodynamic assessment of hydrothermal alkalifeldspar – mica – aluminosilicate equilibria. Geochim.Cosmochim. Acta 55, 989-1004.

TILLEY, C.E. (1929): On larnite (calcium orthosilicate, a newmineral) and its associated minerals from the limestonecontact-zone of Scawt Hill, Co. Antrim. Mineral. Mag. 22,77-86.

________ (1948): The gabbro – limestone contact zone ofCamas Mor, Muck, Invernessshire. Bull. Comm. Géol.Finlande 140, 97-104.

1405 39#5-oct-01-2277-15 26/10/01, 13:021433


________ (1951): A note on the progressive metamorphism ofsiliceous limestones and dolomites. Geol. Mag. 88, 175-178.

UCHIDA, E. & IIYAMA, J.T. (1981): On kamaishilite, Ca2Al2SiO6OH)2, a new mineral dimorphous with bicchulite fromthe Kamaishi mine, Japan. Proc. Japan. Acad. 57, ser. B,239-243.

WIECHMANN, M.J. (1995): Contact Metamorphism and SkarnFormation at Crestmore, California. Ph.D. thesis, JohnsHopkins Univ., Baltimore, Maryland.

WILLEMSE, J. & BENSCH, J.J. (1964): Inclusions of originalcarbonate rocks in gabbro and norite of the eastern part ofthe Bushveld complex. Trans. Geol. Soc. S. Afr. 67, 1-87.

Received January 6, 2000, revised manuscript acceptedAugust 31, 2001.

APPENDIX : CALCITE SOLUBILITY IN A CHLORIDE-BEARING FLUID

Calcite is dissolved in a chloride fluid according toreaction (a):

CaCO3 + 2 HCl = CaCl2 + CO2 + H2O (a)

CaCl2 is not appreciably dissociated at the temperatureand pressure conditions of formation estimated for theskarn. However, owing to the possible contribution ofCa(OH)2, consideration of this reaction only provides aminimum estimation of the solubility of calcite. Owingto the low density of the fluid, the equations of state foraqueous species do not allow us to calculate the equi-librium constant, but an order of magnitude can be ob-tained from the experimental study of the equilibriumbetween wollastonite, quartz and chloride fluid (Gunter& Eugster 1978):

CaSiO3 + 2 HCl° = CaCl2° + SiO2 + H2O (b)

The corresponding equilibrium constant, Kb, is ex-pressed by log Kb = log a(CaCl2°) – 2 log a(HCl°) + log

a(H2O), and measured as log Kb ≈ log m(CaCl2) – 2 logm(HCl ) + log a(H2O) = 1.78 (1 kilobar) and 2.49 (2 ki-lobars), assuming that CaCl2° was the main aqueousspecies of Ca and conventionally taking the activitycoefficients of uncharged aqueous species as 1.

Under the same assumptions, the constant of equi-librium (a) is calculated from Kb and the thermodynamicdata for wollastonite, quartz, calcite and CO2 from theTHERMOCALC database: log Ka = log a(CaCl2°) – 2log a(HCl°) + log a(H2O) + log f(CO2) ≈ 5.40 (1 kilo-bar) and 6.22 (2 kilobars).

If the reaction �V remained constant between 0.75kilobar and 2 kilobars, log Ka would be 5.2 at 0.75 kbar,and the ratio a(CaCl2°)/a(HCl°)2 at equilibrium withcalcite would be about 103.8–4.0 for CO2 pressure in therange 15 – 30 bars. Although the assumption of con-stant �V is probably a poor one, the ratio m(CaCl2)/m(HCl)2 is large enough to ensure that the initial HClcontent of the fluid would be quantitatively convertedto CaCl2 upon reaction with calcite.

1405 39#5-oct-01-2277-15 26/10/01, 13:021434

THE MELILITE-BEARING HIGH-TEMPERATURE SKARNS …rruff.info/doclib/cm/vol39/CM39_1405.pdf · MELILITE-BEARING HIGH-TEMPERATURE SKARNS, ROMANIA 1407 (mainly granite porphyry and granophyre),

Documents